+

CN107730087A - Forecast model training method, data monitoring method, device, equipment and medium - Google Patents

Forecast model training method, data monitoring method, device, equipment and medium Download PDF

Info

Publication number
CN107730087A
CN107730087A CN201710853244.8A CN201710853244A CN107730087A CN 107730087 A CN107730087 A CN 107730087A CN 201710853244 A CN201710853244 A CN 201710853244A CN 107730087 A CN107730087 A CN 107730087A
Authority
CN
China
Prior art keywords
service data
data
training
monitoring
model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201710853244.8A
Other languages
Chinese (zh)
Inventor
张川
顾青山
金鑫
温善安
李磊
王定鑫
李泳章
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Priority to CN201710853244.8A priority Critical patent/CN107730087A/en
Priority to PCT/CN2017/108533 priority patent/WO2019056499A1/en
Publication of CN107730087A publication Critical patent/CN107730087A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0202—Market predictions or forecasting for commercial activities
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Human Resources & Organizations (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Economics (AREA)
  • Artificial Intelligence (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Software Systems (AREA)
  • Game Theory and Decision Science (AREA)
  • Evolutionary Biology (AREA)
  • Biomedical Technology (AREA)
  • Educational Administration (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Technology Law (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The present invention discloses a kind of forecast model training method, data monitoring method, device, equipment and medium.The forecast model training method includes：Original service data are carried out with time-labeling and is divided by the default time limit, obtains the training business datum for carrying time sequence status；The training business datum is divided into training set and test set by preset ratio；Long short-term memory Recognition with Recurrent Neural Network model is trained using the training set, obtains original predictive model；The original predictive model is tested using the test set, obtains target prediction model.The forecast model training method has the advantages of timing is strong, accuracy rate is high when being predicted.

Description

Prediction model training method, data monitoring method, device, equipment and medium
Technical Field
The invention relates to the field of data monitoring, in particular to a prediction model training method, a data monitoring method, a device, equipment and a medium.
Background
With the development of market economy, competition among enterprises is more and more intense, and in order to improve the competitiveness of the enterprises, the enterprises predict future business development trends by performing big data analysis on historical business data so as to adjust strategies. An SVM (Support Vector Machine) model is adopted to carry out big data analysis on historical service data currently, the mode lacks consideration on self data autocorrelation and does not have the capability of predicting time sequence data, and the accuracy of a data prediction result is low.
Disclosure of Invention
The embodiment of the invention provides a prediction model training method, a data monitoring method, a device, equipment and a medium, and aims to solve the problem of low accuracy of a current data prediction result.
In a first aspect, an embodiment of the present invention provides a predictive model training method, including:
carrying out time marking on original service data and dividing according to a preset time limit to obtain training service data carrying a time sequence state;
dividing the training service data into a training set and a test set according to a preset proportion;
training the training set by adopting a long-time memory cyclic neural network model to obtain an original prediction model;
and testing the original prediction model by adopting the test set to obtain a target prediction model.
In a second aspect, an embodiment of the present invention provides a data monitoring method, including:
acquiring a data monitoring instruction, wherein the data monitoring instruction comprises current time, a preset time limit and a monitoring index;
acquiring monitoring service data based on the data monitoring instruction, wherein the monitoring service data is historical service data relative to the monitoring index within the preset time limit before the current time;
and predicting the monitoring service data by adopting the target prediction model to obtain predicted service data.
In a third aspect, an embodiment of the present invention provides a prediction model training apparatus, including:
the training service data acquisition module is used for carrying out time marking on the original service data and dividing the original service data according to a preset time limit to acquire training service data carrying a time sequence state;
the data dividing module is used for dividing the training service data into a training set and a test set according to a preset proportion;
the original prediction model acquisition module is used for training the long-time memory cyclic neural network model by adopting the training set to acquire an original prediction model;
and the target prediction model acquisition module is used for testing the original prediction model by adopting the test set to acquire a target prediction model.
In a fourth aspect, an embodiment of the present invention provides a data monitoring apparatus, including:
the data monitoring instruction acquisition module is used for acquiring a data monitoring instruction, and the data monitoring instruction comprises current time, a preset time limit and a monitoring index;
a monitoring service data obtaining module, configured to obtain monitoring service data based on the data monitoring instruction, where the monitoring service data is specifically historical service data, which is within the preset time limit before the current time and is relative to the monitoring index;
and the predicted service data acquisition module is used for predicting the monitoring service data by adopting the target prediction model to acquire predicted service data.
In a fifth aspect, an embodiment of the present invention provides a terminal device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the predictive model training method when executing the computer program.
In a sixth aspect, an embodiment of the present invention provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the steps of the data monitoring method are implemented.
In the prediction model training method, device, equipment and medium provided by the embodiment of the invention, the training service data carrying a time sequence state is obtained by performing time marking on the obtained original service data and dividing according to a preset time limit, so that the time sequence of the training service data is increased, and the accuracy of the prediction model is improved. And then training by adopting a long-time memory cyclic neural network model to obtain an original prediction model, so that the training of the original prediction model has the advantages of high efficiency and high accuracy. And finally, testing the original prediction model by adopting the training service data in the test set to obtain a target prediction model so as to improve the accuracy of prediction of the target prediction model, and enabling the obtained target prediction model to have time sequence due to the fact that the training service data have time sequence.
In the data monitoring method, the data monitoring device, the data monitoring equipment and the data monitoring medium, the target prediction model is adopted to monitor historical service data, namely monitoring service data, corresponding to the monitoring index within the preset time limit before the current time, and the method has the advantages of strong time sequence and high accuracy, so that the time sequence and the accuracy of the data monitoring result are improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.
Fig. 1 is a flowchart of a predictive model training method provided in embodiment 1 of the present invention.
Fig. 2 is a specific schematic diagram of step S130 in fig. 1.
Fig. 3 is a specific diagram of step S131 in fig. 2.
Fig. 4 is a schematic diagram of the prediction model training apparatus provided in embodiment 2 of the present invention.
Fig. 5 is a flowchart of a data monitoring method provided in embodiment 3 of the present invention.
Fig. 6 is another flowchart of the data monitoring method provided in embodiment 3 of the present invention.
Fig. 7 is another flowchart of the data monitoring method provided in embodiment 3 of the present invention.
Fig. 8 is a schematic diagram of a data monitoring apparatus provided in embodiment 4 of the present invention.
Fig. 9 is a schematic diagram of a terminal device provided in embodiment 6 of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
Fig. 1 shows a flowchart of a predictive model training method in the present embodiment. The prediction model training method is applied to terminal equipment configured by financial institutions such as banks, securities and insurance institutions or other institutions, and is used for performing prediction model training by using business data generated by the financial institutions or other institutions so as to predict future business data based on the trained prediction model and achieve the purpose of artificial intelligent monitoring. As shown in fig. 1, the predictive model training method includes the following steps:
s110: and carrying out time marking on the original service data and dividing according to a preset time limit to obtain training service data carrying a time sequence state.
The original business data is business data formed in the production and operation process in a financial institution or other institutions. The business data includes, but is not limited to, sales, billing amount, employee attendance, customer number and campaign amount in this embodiment. Specifically, the time marking of the original service data refers to that the original service data carries a time tag, and the time tag may be the time of the day when the original service data is formed. If the original service data is sales of 6 month 7, the time stamp of the original service data is 6 month 7 days.
In this embodiment, a Hadoop big data platform is used to collect original business data formed by financial institutions or other institutions every day, and time marking is performed on the counted original business data. The Hadoop big data platform enables a user to develop a distributed program without knowing details of a distributed bottom layer, and high-speed operation and storage are carried out, so that the acquisition efficiency of original business data is improved. The Hadoop refers to a Distributed System infrastructure, and implements a Distributed File System (HDFS) by Hadoop. The HDFS has the characteristic of high fault tolerance, is designed to be deployed on cheap hardware, can provide high throughput to access data of an application program, is suitable for the application program with an ultra-large data set, and has the advantage of high acquisition efficiency by adopting a Hadoop big data platform to adopt original service data.
Specifically, the preset time limit is a preset time limit for dividing the training service data with the time sequence state, and the preset time limit may be 1 month, 2 months, half a year, or other time limit values. If the preset time limit is 1 month, when time labeling is performed on the original service data and the original service data are divided according to the preset time limit, the original service data in 1 continuous month in the original service data carrying the time tag formed after the time labeling is performed on the original service data are required to be used as training service data carrying a time sequence state. That is, if the preset time limit is 1 month, when the prediction model to be trained is to predict future service data 1-N days after the preset time limit by using the original service data within the preset time limit, the training service data carrying the time sequence state formed by the prediction model can be represented in the form of (the original service data within the preset time limit, the original service data 1-N days after the preset time limit). Wherein N is an integer greater than 1. For example, in this embodiment, (original service data No. 6-30, original service data No. 7-1), (original service data No. 6-7-1, and original service data No. 7-2) and the like may be used as training service data carrying a time sequence state.
S120: and dividing the training service data into a training set and a test set according to a preset proportion.
The preset proportion is preset and is used for classifying the training service data. The preset ratio may be a ratio obtained from historical experience. The training set is a learning sample data set, and a classifier is established by matching some parameters, that is, training a machine learning model by using training service data in the training set to determine parameters of the machine learning model. The test set (test set) is used to test the resolving power, such as recognition rate, of the trained machine learning model. In this embodiment, the training service data may be classified according to a ratio of 9:1, that is, 90% of the training service data may be used as a training set, and the remaining 10% may be used as a test set.
S130: and training the training set by adopting a long-time memory cyclic neural network model to obtain an original prediction model.
The long-short term memory (LSTM) model is a time recurrent neural network model, and is suitable for processing and predicting important events with time series and relatively long time series intervals and delays. The LSTM model has a time memory function and is therefore used to process training traffic data carrying timing states. The LSTM model is one of neural network models with long-term memory capability, and has a three-layer network structure of an input layer, a hidden layer, and an output layer. The input layer is the first layer of the LSTM model, and is used to receive external signals, that is, is responsible for receiving training service data carrying a timing sequence state in a training set. The output layer is the last layer of the LSTM model and is used for outputting signals to the outside, i.e. responsible for outputting the calculation results of the LSTM model. The hidden layer is a layer except the input layer and the output layer in the LSTM model and is used for processing training service data in a training set and obtaining a calculation result of the LSTM model. The original prediction model is obtained by adopting an LSTM model to carry out multiple iterations on training service data carrying a time sequence state in a training set and verifying the training service data. Understandably, the model training using the LSTM model increases the timing sequence of the training business data, thereby improving the accuracy of the prediction model. In this embodiment, the output layer of the LSTM model performs regression processing using Softmax (regression model) for classifying and outputting the weight matrix. The Softmax (regression model) is a classification function commonly used in a neural network, and maps the output of a plurality of neurons into a [0,1] interval, so that the probability can be understood, the calculation is simple and convenient, and the multi-classification output is performed, and the output result is more accurate.
It can be understood that, when the long-term and short-term memory recurrent neural network model is adopted to train the training set to obtain the original prediction model, part of the training service data carrying the time sequence state in the training set can be further divided into verification sets (validations sets) according to a certain proportion. The verification set can adjust parameters of the classifier of the learned model, and determine parameters of a network structure or a complex program controlling the model. In this implementation, the training service data in the original training set is calculated according to the following formula 8: 1, classifying the training service data, so that the proportion of the training service data in the training set, the verification set and the test set is 8: 1: 1.
in a specific embodiment, as shown in fig. 2, in step S130, training a training set by using a long-and-short-term memory recurrent neural network model to obtain an original prediction model, specifically including the following steps:
s131: and training the training set by adopting a time-based back propagation algorithm to obtain an original state model.
The Back Propagation Through Time (BPTT) algorithm is an optimization algorithm applied to a neural network model based on Time series data, the BPTT algorithm is a very intuitive method, a propagation path is a circular path, and parameters on the propagation path are sharable parameters. The BPTT algorithm includes a forward Propagation (forward Propagation) algorithm and a Back Propagation (Back Propagation) algorithm. Wherein, the forward propagation (forward propagation) algorithm is processed according to the time sequence; and the Back Propagation (Back Propagation) algorithm is to transfer the accumulated residual Back from the last time and train the neural network model. The original state model refers to a model obtained through multiple iterations of the BPTT algorithm. In this embodiment, the BPTT algorithm is applied to the LSTM model to train the training set, specifically, a Forward Propagation (Forward Propagation) algorithm and a backward Propagation (back Propagation) algorithm are used to process the training service data in the training set to obtain the original state model, and the BPTT algorithm has the advantages of strong time sequence and high accuracy.
In a specific embodiment, as shown in fig. 3, in step S131, training a training set by using a time-based back propagation algorithm to obtain an original state model, specifically including the following steps:
s1311: training the training set by adopting a forward propagation algorithm to obtain a first state parameter of the original state model; the forward propagation algorithm has the formulaAndwherein S istAn output representing a hidden layer at the current time;representing the weight from the previous moment to the current moment of the hidden layer;representing the weight from the input layer to the output layer;a prediction output representing a current time;representing weights of the hidden layer to said output layer.
The training set is trained by adopting a Forward Propagation (Forward Propagation) algorithm, which means that training service data are trained by adopting a Forward Propagation calculation formula according to the sequence of time sequence states carried by the training service data in the training set. The first state parameter refers to a parameter obtained in an initial iteration process when model training is performed based on training service data.
As will be appreciated, the forward propagation algorithm is to take the input X of the current time instanttAnd the output S of the hidden unit at the previous momentt-1I.e. the output S of the memory cells in the hidden layer in the LSTM modelt-1As the input of the hidden layer, the output S of the hidden layer at the current time is obtained by the transformation of the activation function tanh (hyperbolic tangent)tThe predicted output at time t is usedIt is shown that U represents a weight from a previous time to a current time of the hidden layer, W represents a weight from the input layer to the hidden layer, and V represents a weight from the hidden layer to the output layer. From this, the predicted outputOutput S from the current timetCorrelation, StThe input at the time t and the state at the time t-1 are included, so that all information on a time sequence is reserved in prediction output, and the prediction output has time sequence. Because the expression capability of the linear model is not enough, in the embodiment, tanh (hyperbolic tangent) is used as an activation function, and a nonlinear factor can be added, so that the trained original prediction model can solve a more complex problem. Moreover, the activation function tanh (hyperbolic tangent) has the advantage of high convergence rate, so that the training time can be saved, and the training efficiency can be improved.
S1312: error calculation is carried out on the first state parameter by adopting a back propagation algorithm, and a second state parameter of the original state model is obtained; the calculation formula of the back propagation algorithm isWherein,a prediction output representing time t; otRepresents time t andthe corresponding true value.
Specifically, the error calculation of the first state parameter by using a Back Propagation (Back Propagation) algorithm means that the optimization parameters, i.e., the three weight parameters U, V and W in the present embodiment, are updated in a time-reversed order. In this embodiment, the error calculation is performed by defining a loss function at the t-th time of backward propagation as cross entropy, that is, a formula is adoptedAnd (6) performing calculation. Finally, calculating the partial derivative of each layer according to a chain type derivative methodThe three weight parameters, U, V and W, are updated based on the three rates of change to obtain adjusted state parameters. Wherein, therefore, the three change rates can be obtained by calculating partial derivatives of the loss function at each moment and then adding the partial derivatives, so as to update the weight parameters. Because the gradient can exponentially increase along with the increasing of the number of the backward propagation layers to cause the phenomenon of gradient disappearance, the problem of gradient disappearance can be well solved by matching the cross entropy loss function with the tanh activation function in the embodiment, and the training accuracy is increased.
S1313: and acquiring an original state model based on the second state parameter.
The original state model is a model formed by converging the second state parameter after multiple iterations based on the second state parameter acquired in step S1312. In this embodiment, in the training process of steps S1311 to S1313, the problem of gradient disappearance can be solved well by matching the cross entropy loss function with the activation function tanh (hyperbolic tangent), and the accuracy of training is increased; moreover, the activation function tanh (hyperbolic tangent) has the advantage of high convergence rate, so that the training time can be saved, and the training efficiency can be improved.
S132: and (5) inspecting the original state model by adopting an R-square method to obtain an original prediction model.
Wherein R-square is called the determination coefficient of the equation, and the closer to 1 the R-square is between 0 and 1, the stronger the interpretability of the variable of the equation on y is. In this embodiment, the original state model obtained in step S1313 is checked by using the training service data in the verification set as a check sample. It can be understood that the original state model is obtained by checking the fitting degree of the original state model through the R-square method by using the training service data in the verification set, so as to determine the structure of the model.
Specifically, the formula of the R-square method is, in the form of a regression sum of squares,is the sum of the squares of the total deviations. It will be appreciated that the size of the R-square determines how well the model fits, with a larger R-square indicating a better fit. In the embodiment, in each iterative operation process, an R-square value is obtained, the R-square value is compared with a preset fitting value, and if the R-square value is greater than the preset fitting value, it is indicated that the original state model is well fitted, that is, the convergence effect is good, and the original state model can be used as an original prediction target.
In the embodiment, the training service data in the training set is trained by adopting a time-based backward propagation algorithm, the forward propagation algorithm processing process has the advantages of strong time sequence and high convergence rate, and the backward propagation algorithm processing process can enable the prediction model training to have the advantages of high efficiency and high accuracy. And the R-square method is adopted to further verify the original state model by utilizing the training service data in the verification set, so that the accuracy of the original prediction model is improved.
S140: and testing the original prediction model by adopting a test set to obtain a target prediction model.
In this embodiment, the training service data carrying the time sequence state in the test set can be represented in the form of (original service data within a preset time limit, original service data 1-N days after the preset time limit), and the original service data within the preset time limit in each training service data in the test set is input into the original prediction model for testing to obtain original prediction data; comparing the original prediction data with original service data which is 1-N days after the real preset time limit in the training service data to obtain the difference value of the original prediction data and the original service data, and judging whether the difference value is within the preset range; if the difference value is within a preset range, determining that the prediction result is accurate; otherwise, if the difference is not within the preset range, the prediction result is determined to be inaccurate. The preset range is a range value used for evaluating whether the prediction result is accurate.
As can be understood, in step S140, all the training service data in the test set are input into the original prediction model for testing, and the probability that the prediction result is accurate is obtained (i.e., the number of accurate prediction results is divided by the number of all the training service data in the training set). Judging whether the probability of the accurate prediction result is greater than a preset probability, if so, determining that the original prediction model is more accurate to take the original prediction model as a target prediction model; otherwise, if the probability of the accuracy of the prediction result is not greater than the preset probability, the original prediction result is determined to be not accurate enough, and the test is performed again after the training in steps S110-S130 is performed.
In the prediction model training method provided in this embodiment, the training service data carrying a time sequence state is obtained by time-labeling the obtained original service data and dividing the obtained original service data according to a preset time limit, so that the time sequence of the training service data is increased, and the accuracy of the prediction model is improved. And then training by adopting a long-time memory cyclic neural network model to obtain an original prediction model, so that the training of the original prediction model has the advantages of high efficiency and high accuracy. And finally, testing the original prediction model by adopting the training service data in the test set to obtain a target prediction model so as to improve the accuracy of prediction of the target prediction model, and enabling the obtained target prediction model to have time sequence due to the fact that the training service data have time sequence.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
Example 2
Fig. 4 is a schematic block diagram of a predictive model training apparatus corresponding to the predictive model training method in embodiment 1. As shown in fig. 4, the prediction model training apparatus includes a training business data obtaining module 110, a data partitioning module 120, an original prediction model obtaining module 130, and a target prediction model obtaining module 140. The implementation functions of the training service data obtaining module 110, the data dividing module 120, the original prediction model obtaining module 130, and the target prediction model obtaining module 140 correspond to the steps corresponding to the prediction model training method in embodiment 1 one to one, and for avoiding redundancy, detailed description is not needed in this embodiment.
The training service data obtaining module 110 is configured to perform time marking on the original service data and divide the original service data according to a preset time limit to obtain training service data carrying a time sequence state.
The data dividing module 120 is configured to divide the training service data into a training set and a test set according to a preset ratio.
And an original prediction model obtaining module 130, configured to train the training set by using a long-term and short-term memory recurrent neural network model, so as to obtain an original prediction model.
Preferably, the raw prediction model acquisition module 130 includes a raw state model acquisition unit 131 and a raw prediction model acquisition unit 132.
The original state model obtaining unit 131 is configured to train the training set by using a time-based back propagation algorithm to obtain an original state model.
And an original prediction model obtaining unit 132, configured to examine the original state model by using an R-square method, and obtain an original prediction model.
Preferably, the raw state model obtaining unit 131 includes a first state parameter obtaining sub-unit 1311, a second state parameter obtaining sub-unit 1312, and a raw state model obtaining sub-unit 1313.
A first state parameter obtaining subunit 1311, configured to train the training set by using a forward propagation algorithm, and obtain a first state parameter of the original state model.
The second state parameter obtaining subunit 1312 is configured to perform error calculation on the first state parameter by using a back propagation algorithm, and obtain a second state parameter of the original state model.
An original state model obtaining subunit 1313, configured to obtain an original state model based on the second state parameter.
And the target prediction model obtaining module 140 is configured to test the original prediction model by using a test set to obtain a target prediction model.
Example 3
Fig. 5 shows a data monitoring method in the present embodiment. The data monitoring method can be applied to terminal equipment configured by financial institutions such as banks, securities and insurance, or other institutions needing data monitoring, so that the monitoring of future service data is realized based on the acquired original service data, and an adjustment strategy is made in time. As shown in fig. N, the data monitoring method includes the following steps:
s210: and acquiring a data monitoring instruction, wherein the data monitoring instruction comprises the current time, a preset time limit and a monitoring index.
The data monitoring instruction is an instruction for controlling the terminal device to perform service data monitoring. The current time is the system time of the terminal equipment, and the current time is specific to the current system. The monitoring index refers to the type of the service data to be monitored, and the monitoring index can correspond to a specific service data. Thus, monitoring metrics include, but are not limited to, sales, billing amounts, clerk attendance, customer number and campaign volume. The preset time limit is a preset time limit, and the preset time limit is used for determining historical service data corresponding to the monitoring prediction. The preset time limit in this embodiment is the same as the preset time limit in embodiment 1, so as to ensure that the target prediction model obtained by training in embodiment 1 can be applied to service data monitoring in this embodiment, and improve the accuracy of data prediction.
S220: and acquiring monitoring service data based on the data monitoring instruction, wherein the monitoring service data is specifically historical service data relative to the monitoring index within a preset time limit before the current time.
The historical business data is the business data formed by the financial institution or other institutions in the production and operation process before the current time, and is stored in the Hadoop big data platform. After the terminal device obtains the data monitoring instruction, all historical service data corresponding to the monitoring index can be obtained based on the monitoring index in the data monitoring instruction, and the historical service data which is before the current time and within the preset time limit is selected from all the historical service data to serve as the monitoring service data.
Specifically, the monitoring service data is historical service data corresponding to the monitoring index within a preset time limit before the current time, which indicates that the monitoring service data is determined in a "rolling" manner in this embodiment. If the preset time limit is T and the current time is the predicted first day, taking historical service data in the past T time before the predicted first day as monitoring service data; if the current time is the predicted second day, taking the historical service data of the predicted first day and the historical service data of the past T-1 day as monitoring service data; if the preset period is 30 days and the current time is 7 months and 1 day, the monitoring service data is historical service data of 30 days from 6 months and 1 day to 6 months and 30 days, and the service data to be predicted is data of 7 months and 1 day or after 7 months and 1 day. The rolling mode determines the monitoring service data, so that the time sequence of the monitoring service data is stronger, and the accuracy of the monitoring result is increased.
S230: and predicting the monitoring service data by adopting a target prediction model to obtain predicted service data.
The target prediction model is obtained by using the prediction model training method in embodiment 1, and has the advantages of good time sequence, high prediction efficiency and high accuracy, so that when the target prediction model is used for predicting the monitored business data in step S230, more accurate predicted business data can be quickly obtained, and an enterprise can know possible business data developed according to the current trend based on the predicted business data so as to perform decision adjustment.
In a specific embodiment, as shown in fig. 6, in the data monitoring method, after step S230, the following steps are further included:
s240: and acquiring standard service data corresponding to the monitoring index.
The standard business data refers to business data preset by an enterprise, and the standard business data is applied to enterprise annual planning. For example, when an enterprise makes annual planning, monthly planning needs to be performed on business data such as sales per month, billing amount, attendance rate of salesmen, number of customers to be received, propaganda activity amount and the like, and the monthly business data is determined as standard business data. Or when the enterprise makes monthly plans, the enterprise needs to plan business data such as daily sales, billing amount, attendance rate of the salesmen, the number of customers to be received, publicity activity amount and the like, and determine daily standard business data. It is understood that the standard service data may be a standard value or a standard range.
S250: and judging whether the predicted service data accords with the standard service data or not based on the predicted service data and the standard service data.
The predicted service data is data obtained by predicting the monitored service data by using the target prediction model in step S230, and is a predicted value. When judging whether the predicted service data accords with the standard service data, if the standard service data is a specific standard value, judging whether the predicted service data is larger than or smaller than the standard service data so as to determine whether the predicted service data accords with the standard service data. If the predicted service data is sales volume, when the predicted service data is greater than the standard service data, the predicted service data conforms to the standard service data; otherwise, when the predicted service data is not larger than the standard service data, the predicted service data does not conform to the standard service data, and monitoring personnel need to be reminded. If the predicted service data is the customer complaint volume, when the predicted service data is smaller than the standard service data, the measured service data conforms to the standard service data; on the contrary, if the predicted service data is not less than the standard service data, the measured service data does not conform to the standard service data, and monitoring personnel need to be reminded. If the standard service data is in a standard range, when the predicted service data is in the standard range, the predicted service data is determined to be in accordance with the standard service data; and otherwise, when the predicted service data is not in the standard range, determining that the predicted service data does not conform to the standard service data.
S260: and if the predicted service data do not accord with the standard service data, acquiring a monitoring result of abnormal monitoring data.
It can be understood that when it is determined that the predicted business data does not meet the standard business data according to step S250, a monitoring result of monitoring data abnormality is obtained, so that monitoring personnel of the enterprise make decision adjustment based on the monitoring result.
In the data monitoring method provided in this embodiment, through steps S240 to S260, the obtained predicted service data is compared with standard service data preset by a user to determine whether a monitoring result indicating that the monitoring data is abnormal exists, so as to remind an enterprise internal manager to perform decision adjustment based on the monitoring result. Moreover, the predicted business data is obtained after the target prediction model obtained in embodiment 1 is adopted to predict the monitoring business data, so that more accurate predicted business data can be rapidly obtained, and an enterprise can know possible business data developed according to the current trend based on the predicted business data so as to perform decision adjustment.
Further, in this embodiment, the data monitoring instruction acquired in step S210 may further include a monitoring mailbox. When the terminal device executes steps S210-S230 to obtain the predicted service data, the predicted service data may be sent to the monitoring mailbox, so that a user of the monitoring mailbox can know the current sending trend of the enterprise according to the predicted service data, and make a decision adjustment accordingly. Or, when the terminal device executes steps S240-S260 to obtain a monitoring result of monitoring data abnormality, the monitoring result is sent to the monitoring mailbox, so that the use of the monitoring mailbox can know whether standard business data pre-made by the enterprise can be realized according to the current development trend of the enterprise according to the monitoring result, and thus, whether decision adjustment is needed is determined, and thus, the production and operation activities of the enterprise are optimized.
Further, in this embodiment, the data monitoring instruction obtained in step S210 is specifically a real-time monitoring instruction, and the timing monitoring instruction not only includes the current time, the preset time limit and the monitoring index, but also includes a trigger time point and a monitoring mailbox. The trigger time point is a time point for triggering the terminal device to perform steps S220-S230 or to perform steps S230-S260. For example, the triggering time point in the timing monitoring instruction may be set to 1 pm per day, so as to control the terminal device to perform steps S220-S230 or perform steps S230-S260 at 1 pm per day. When the terminal device executes the steps of steps S220-S230 to obtain the predicted service data based on the timing monitoring instruction, the predicted service data can be sent to the monitoring mailbox, so that a user of the monitoring mailbox can know the current sending trend of the enterprise according to the predicted service data, and make a decision adjustment accordingly. Or, when the terminal device executes the monitoring result of the abnormal monitoring data obtained in the steps S240 to S260 based on the timing monitoring instruction, the monitoring result is sent to the monitoring mailbox, so that the use of the monitoring mailbox can know whether the standard business data preset by the enterprise can be realized according to the current development trend of the enterprise according to the monitoring result, and thus whether decision adjustment is needed or not is determined, and the production and operation activities of the enterprise are optimized.
In a specific embodiment, steps S210 to S230 may obtain predicted service data, where the predicted service data corresponds to a target time, and the target time is a time corresponding to the predicted service data. The predicted business data can be one of sales amount, bill amount, attendance rate of the salesman, number of customers to be received, propaganda activity amount and the like. Therefore, steps S210-S230 may be repeatedly performed to obtain at least two predicted traffic data for the same target time. As shown in fig. 7, the data monitoring method provided in this embodiment obtains at least two predicted service data at the same target time, and then further includes the following steps:
s231: and acquiring the weight corresponding to each predicted service data.
In this embodiment, a linear regression algorithm is used to perform regression processing on historical service data in advance, and a weight of predicted service data corresponding to the historical service data is obtained. Wherein the calculation formula of the linear regression algorithm is hθ(x)＝θ0+θ1x1+θ2x2+…+θnxnWherein h isθ(x) For the hypothesis function, each θ is an angle vector between input values, and each x is a corresponding plurality of variable vectors, i.e., a value corresponding to each type of historical service data. Then, a Cost Function (Cost Function) is constructed, and if the Cost Function is smaller, the fitting degree of the linear regression is better. The expression of the Cost Function (Cost Function) is as follows：Wherein x is(i)Representing the ith element in the vector x, namely the ith service data; y is(i)Representing the ith element in the vector y, namely the sales corresponding to the ith service data; h isθ(x(i)) Representing a known hypothetical function, and m is the number of classes of traffic data. Then, the minimum value of the cost function (cost function) is found out according to a gradient descent method, the step size of the next step is determined, and an initial value theta is given arbitrarily0,θ1Determining a downward direction, and moving down a predetermined step, and updating theta0,θ1And stopping the descent when the height of the descent is less than a certain defined value. The expression of the gradient descent method isAnd acquiring the weight corresponding to each historical service data by calculating the value of each theta, namely acquiring the weight corresponding to each predicted service data.
S232: processing at least two pieces of predicted service data and corresponding weights by adopting a weighted operation algorithm to obtain a predicted index value of a target time; the weighting algorithm is Pi＝Σ(±viwi) Wherein P isiTo predict the index value, viTo predict traffic data, wiIs the weight of each predicted traffic data.
The prediction index value is obtained by performing weighting operation on at least two pieces of predicted service data. It is understood that the prediction index value is an evaluation value obtained by comprehensively evaluating at least two predicted service data. In the process of determining the prediction index data, part of the prediction service data plays a positive role on the prediction index value (the larger the value is, the better the value is), and the part of the prediction service data needs to be positive by the value of each weight in calculation; on the contrary, the more reverse the part of the predicted service data is on the value of the prediction index (the smaller the value is, the better the value is, such as the customer complaint amount, etc.), the calculation needs to make the part of the predicted service data take the negative value of the respective weight.
S233: and acquiring a target index value corresponding to the target time.
The target index value is an index value for comprehensively evaluating at least two service data preset by an enterprise. In this embodiment, the target index value is a standard value obtained by using the above weighting algorithm according to at least two standard service data, and the target index value is stored in a database connected to the terminal device, so that the target index value is directly called based on a target time period when used, and the efficiency of obtaining the target index value is improved.
It is understood that steps S231 to S232 and step S233 have no execution sequence restriction, and steps S231 to S232 may be executed first and then step S233 is executed, or step S233 may be executed first and then steps S231 to S232 are executed.
S234: and judging whether the prediction index value meets the target index value.
That is, the prediction index value obtained in step S232 and the target index value obtained in step S233 are determined, and whether or not the prediction index value matches the target index value is determined. In this embodiment, when the predicted index value and the target index value are both a specific value, if the predicted index value is greater than the target index value, the predicted index value is determined to conform to the target index value; otherwise, if the prediction index value is not greater than the target index value, the prediction index value is determined not to conform to the target index value. Or, when the prediction index value and the target index value are both in the data range, if the prediction index value is in the data range corresponding to the target index value, the prediction index value is determined to be in accordance with the target index value; on the contrary, if the prediction index value is not in the data range corresponding to the target index value, the prediction index value is determined not to conform to the target index value.
S235: and if the predicted index value does not accord with the target index value, acquiring reminding information.
It is to be understood that when it is determined that the predicted index value does not meet the target index value according to step S234, a reminding message is obtained, so that the monitoring personnel of the enterprise makes a decision adjustment based on the reminding message. Further, in this embodiment, the data monitoring instruction obtained in step S210 may further include monitoring a mailbox or a contact phone; after the reminding information is obtained by executing the steps S231-S235, the reminding information is sent to a monitoring mailbox or a contact phone, so that relevant personnel can know the current development trend of the enterprise according to the reminding information, and further determine whether decision adjustment is needed or not.
In the data monitoring method provided by this embodiment, through steps S231-S235, relevant personnel of the enterprise can learn the current development trend of the enterprise from a macro level according to the acquired reminding information, so that the enterprise can adjust the policy in time, and further optimize the production and operation activities of the enterprise.
Example 4
Fig. 8 is a schematic block diagram of a data monitoring apparatus corresponding to the data monitoring method in embodiment 3 one to one. As shown in fig. 8, the data monitoring apparatus includes a data monitoring instruction obtaining module 210, a monitoring service data obtaining module 220, a predicted service data obtaining module 230, a standard service data obtaining module 240, a predicted service data judging module 250, and a monitoring result obtaining module 260. Moreover, the data monitoring apparatus further includes a weight obtaining module 231, a prediction index value obtaining module 232, a target index value obtaining module 233, a prediction index value determining module 234, and a reminder information obtaining module 235. The implementation functions of the data monitoring instruction obtaining module 210, the monitoring service data obtaining module 220, the predicted service data obtaining module 230, the standard service data obtaining module 240, the predicted service data judging module 250, and the monitoring result obtaining module 260 correspond to the steps corresponding to the data monitoring method in embodiment 3 one to one, and for avoiding repeated descriptions, detailed descriptions are not needed in this embodiment.
The data monitoring instruction obtaining module 210 is configured to obtain a data monitoring instruction, where the data monitoring instruction includes a current time, a preset time limit, and a monitoring index.
The monitoring service data obtaining module 220 is configured to obtain monitoring service data based on the data monitoring instruction, where the monitoring service data is specifically historical service data corresponding to the monitoring index within a preset time limit before the current time.
The predicted service data obtaining module 230 is configured to predict the monitoring service data by using a target prediction model, and obtain predicted service data.
And a standard service data obtaining module 240, configured to obtain standard service data corresponding to the monitoring index.
And a predicted service data judging module 250, configured to judge whether the predicted service data meets the standard service data based on the predicted service data and the standard service data.
And the monitoring result obtaining module 260 is configured to obtain a monitoring result of abnormal monitoring data if the predicted service data does not meet the standard service data.
Preferably, the weight obtaining module 231 is configured to obtain a weight corresponding to each predicted service data.
The prediction index value obtaining module 232 is configured to process the at least two pieces of predicted service data and the corresponding weights by using a weighted operation algorithm to obtain a prediction index value of the target time.
And a target index value obtaining module 233, configured to obtain a target index value corresponding to the target time.
And a prediction index value judgment module 234 for judging whether the prediction index value meets the target index value.
And the reminding information acquisition module 235 is used for acquiring reminding information if the predicted index value does not accord with the target index value.
Example 5
This embodiment provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the method for training a prediction model in embodiment 1 is implemented, and details are not repeated here to avoid repetition. Alternatively, the computer program, when executed by the processor, implements the functions of each module/unit in the prediction model training apparatus in embodiment 2, and is not described herein again to avoid redundancy. Alternatively, the computer program is executed by the processor to implement the functions of the steps in the data monitoring method in embodiment 3, which are not repeated herein to avoid repetition. Alternatively, the computer program is executed by the processor to implement the functions of each module/unit in the data monitoring apparatus in embodiment 4, which is not repeated herein to avoid repetition.
Example 6
Fig. 9 is a schematic diagram of a terminal device according to an embodiment of the present invention. As shown in fig. 9, the terminal device 90 of this embodiment includes: a processor 91, a memory 92 and a computer program 93 stored in the memory 92 and executable on the processor 91. The processor 91, when executing the computer program 93, implements the steps in the various predictive model training method embodiments described above, such as steps S110 to S140 shown in fig. 1. Alternatively, the processor 91, when executing the computer program 93, implements the functions of the modules/units in the above-described device embodiments, such as the functions of the modules 110 to 140 shown in fig. 4. Alternatively, the processor 91 executes the computer program 93 to implement the steps in the above-described respective data monitoring method embodiments, such as the steps S210 to S230 shown in fig. 5. Alternatively, for example, steps S210 to S260 shown in fig. 6. Alternatively, for example, steps S231 to S235 shown in fig. 7. Alternatively, the processor 91, when executing the computer program 93, implements the functions of the modules/units in the above-described device embodiments, such as the functions of the modules 210 to 260 shown in fig. 8.
Illustratively, the computer program 93 may be divided into one or more modules/units, which are stored in the memory 92 and executed by the processor 91 to implement the present invention. One or more of the modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution of the computer program 93 in the terminal device 90. For example, the computer program 93 may be divided into the training service data obtaining module 110, the data dividing module 120, the original prediction model obtaining module 130, and the target prediction model obtaining module 140 in embodiment 2, or the data monitoring instruction obtaining module 210, the monitoring service data obtaining module 220, the prediction service data obtaining module 230, the weight obtaining module 231, the prediction index value obtaining module 232, the target index value obtaining module 233, the prediction index value judging module 234, the reminding information obtaining module 235, the standard service data obtaining module 240, the prediction service data judging module 250, and the monitoring result obtaining module 260 in embodiment 4, and specific functions of each module are as described in embodiment 2 or embodiment 4, which is not described herein.
The terminal device 90 may be a desktop computer, a notebook, a palm computer, a cloud server, or other computing devices. The terminal device may include, but is not limited to, a processor 91, a memory 92. Those skilled in the art will appreciate that fig. 9 is merely an example of a terminal device 90 and does not constitute a limitation of the terminal device 90 and may include more or fewer components than shown, or some components may be combined, or different components, e.g., the terminal device may also include input-output devices, network access devices, buses, etc.
The Processor 91 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The storage 92 may be an internal storage unit of the terminal device 90, such as a hard disk or a memory of the terminal device 90. The memory 92 may also be an external storage device of the terminal device 90, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), or the like provided on the terminal device 90. In particular, the memory 92 may also include both internal and external storage units of the terminal device 90. The memory 92 is used to store computer programs and other programs and data required by the terminal device. The memory 92 may also be used to temporarily store data that has been output or is to be output.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated modules/units, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments may be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain other components which may be suitably increased or decreased as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media which may not include electrical carrier signals and telecommunications signals in accordance with legislation and patent practice.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.

Claims (10)

1. A predictive model training method, comprising:
carrying out time marking on original service data and dividing according to a preset time limit to obtain training service data carrying a time sequence state;
dividing the training service data into a training set and a test set according to a preset proportion;
training the training set by adopting a long-time memory cyclic neural network model to obtain an original prediction model;
and testing the original prediction model by adopting the test set to obtain a target prediction model.
2. The method for training the prediction model according to claim 1, wherein the training set by using the long-time and short-time memory recurrent neural network model to obtain the original prediction model comprises:
training the training set by adopting a time-dependent back propagation algorithm to obtain an original state model;
and (3) inspecting the original state model by adopting an R-square method to obtain the original prediction model.
3. The predictive model training method of claim 2, wherein said training the training set using a back propagation over time algorithm to obtain a raw state model comprises:
training the training set by adopting a forward propagation algorithm to obtain a first state parameter of the original state model; the calculation formula of the forward propagation algorithm isAndwherein S istAn output representing a hidden layer at the current time;representing the weight from the previous moment to the current moment of the hidden layer;representing the weight from the input layer to the output layer;a prediction output representing the current time;representing weights of the hidden layer to the output layer;
adopting a back propagation algorithm to carry out error calculation on the first state parameter to obtain a second state parameter of the original state model; the calculation formula of the back propagation algorithm isWherein,a prediction output representing time t; otRepresents the time t andthe corresponding true value;
and acquiring the original state model based on the second state parameter.
4. A method of data monitoring, comprising:
acquiring a data monitoring instruction, wherein the data monitoring instruction comprises current time, a preset time limit and a monitoring index;
acquiring monitoring service data based on the data monitoring instruction, wherein the monitoring service data is historical service data relative to the monitoring index within the preset time limit before the current time;
predicting the monitoring service data by adopting a target prediction model to obtain predicted service data; wherein the target prediction model is a model obtained by the prediction model training method according to any one of claims 1 to 3.
5. The data monitoring method of claim 4, wherein the predicting the monitored business data by using the target prediction model to obtain predicted business data further comprises:
acquiring standard service data corresponding to the monitoring index;
judging whether the predicted service data accords with the standard service data or not based on the predicted service data and the standard service data;
and if the predicted service data do not accord with the standard service data, acquiring a monitoring result of abnormal monitoring data.
6. The data monitoring method of claim 4, wherein the obtaining of the predicted traffic data comprises: acquiring at least two pieces of predicted service data at the same target time;
the obtaining of the at least two predicted service data at the same target time further includes:
acquiring the weight corresponding to each predicted service data;
processing at least two pieces of predicted service data and corresponding weights by adopting a weighted operation algorithm to obtain a predicted index value of the target time; the weighted arithmetic algorithm is Pi＝Σ(±viwi) Wherein P isiTo predict the index value, viFor the predicted traffic data, wiIs a weight of each of the predicted traffic data;
acquiring a target index value corresponding to the target time;
judging whether the prediction index value meets the target index value;
and if the predicted index value does not accord with the target index value, acquiring reminding information.
7. A predictive model training apparatus, comprising:
the training service data acquisition module is used for carrying out time marking on the original service data and dividing the original service data according to a preset time limit to acquire training service data carrying a time sequence state;
the data dividing module is used for dividing the training service data into a training set and a test set according to a preset proportion;
the original prediction model acquisition module is used for training the long-time memory cyclic neural network model by adopting the training set to acquire an original prediction model;
and the target prediction model acquisition module is used for testing the original prediction model by adopting the test set to acquire a target prediction model.
8. A data monitoring device, comprising:
the data monitoring instruction acquisition module is used for acquiring a data monitoring instruction, and the data monitoring instruction comprises current time, a preset time limit and a monitoring index;
a monitoring service data obtaining module, configured to obtain monitoring service data based on the data monitoring instruction, where the monitoring service data is specifically historical service data, which is within the preset time limit before the current time and is relative to the monitoring index;
a prediction service data obtaining module, configured to use a target prediction model to predict the monitoring service data, so as to obtain prediction service data, where the target prediction model is obtained by using the prediction model training method according to any one of claims 1 to 3.
9. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor when executing the computer program implements the steps of the predictive model training method of any one of claims 1 to 3; alternatively, the processor, when executing the computer program, performs the steps of the data monitoring method according to claims 4 to 6.
10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the predictive model training method according to any one of claims 1 to 3; alternatively, the processor, when executing the computer program, performs the steps of the data monitoring method according to claims 4 to 6.
CN201710853244.8A 2017-09-20 2017-09-20 Forecast model training method, data monitoring method, device, equipment and medium Pending CN107730087A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201710853244.8A CN107730087A (en) 2017-09-20 2017-09-20 Forecast model training method, data monitoring method, device, equipment and medium
PCT/CN2017/108533 WO2019056499A1 (en) 2017-09-20 2017-10-31 Prediction model training method, data monitoring method, apparatuses, device and medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710853244.8A CN107730087A (en) 2017-09-20 2017-09-20 Forecast model training method, data monitoring method, device, equipment and medium

Publications (1)

Publication Number Publication Date
CN107730087A true CN107730087A (en) 2018-02-23

Family

ID=61206686

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710853244.8A Pending CN107730087A (en) 2017-09-20 2017-09-20 Forecast model training method, data monitoring method, device, equipment and medium

Country Status (2)

Country Link
CN (1) CN107730087A (en)
WO (1) WO2019056499A1 (en)

Cited By (73)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108427854A (en) * 2018-03-29 2018-08-21 深圳市智物联网络有限公司 A kind of data analysing method and its relevant device
CN108446864A (en) * 2018-04-10 2018-08-24 广州新科佳都科技有限公司 The fault early warning system and method for Transit Equipment based on big data analysis
CN108446771A (en) * 2018-04-02 2018-08-24 四川长虹电器股份有限公司 A method of preventing Sale Forecasting Model over-fitting
CN108549954A (en) * 2018-03-26 2018-09-18 平安科技（深圳）有限公司 Risk model training method, risk identification method, device, equipment and medium
CN108615360A (en) * 2018-05-08 2018-10-02 东南大学 Transport need based on neural network Evolution Forecast method day by day
CN108615096A (en) * 2018-05-10 2018-10-02 平安科技（深圳）有限公司 Server, the processing method of Financial Time Series and storage medium
CN108764974A (en) * 2018-05-11 2018-11-06 国网电子商务有限公司 A kind of procurement of commodities amount prediction technique and device based on deep learning
CN108764572A (en) * 2018-05-28 2018-11-06 安徽磐众信息科技有限公司 A kind of finance data prediction technique based on Recognition with Recurrent Neural Network
CN108830723A (en) * 2018-04-03 2018-11-16 平安科技（深圳）有限公司 Electronic device, bond yield analysis method and storage medium
CN108846692A (en) * 2018-06-05 2018-11-20 浙江大学城市学院 A kind of consumer spending behavior prediction method based on multifactor Recognition with Recurrent Neural Network
CN108977854A (en) * 2018-08-07 2018-12-11 中铝视拓智能科技有限公司 Electrolyzer temperature monitoring method, device, equipment, system and readable storage medium storing program for executing
CN109034500A (en) * 2018-09-04 2018-12-18 湘潭大学 A kind of mid-term electric load forecasting method of multiple timings collaboration
CN109028480A (en) * 2018-06-08 2018-12-18 杭州古北电子科技有限公司 A kind of thermostatic constant wet control system and its method
CN109190808A (en) * 2018-08-15 2019-01-11 拍拍信数据服务（上海）有限公司 User's behavior prediction method, apparatus, equipment and medium
CN109194423A (en) * 2018-08-13 2019-01-11 中国人民解放军陆军工程大学 Single-frequency point spectrum prediction method based on optimized long-short term memory model
CN109272344A (en) * 2018-08-07 2019-01-25 阿里巴巴集团控股有限公司 Model training method and device, data predication method and device, server
CN109308226A (en) * 2018-08-22 2019-02-05 中国平安人寿保险股份有限公司 Data exception determines method and device, storage medium and electronic equipment
CN109345373A (en) * 2018-09-11 2019-02-15 北京三快在线科技有限公司 Write-off risk early warning method, device, electronic device and computer-readable medium
CN109378065A (en) * 2018-10-30 2019-02-22 医渡云（北京）技术有限公司 Medical data processing method and processing device, storage medium, electronic equipment
CN109660419A (en) * 2018-10-08 2019-04-19 平安科技（深圳）有限公司 Predict method, apparatus, equipment and the storage medium of network equipment exception
CN109672795A (en) * 2018-11-14 2019-04-23 平安科技（深圳）有限公司 Call center resource management method and device, electronic equipment, storage medium
CN109685275A (en) * 2018-12-27 2019-04-26 拉扎斯网络科技（上海）有限公司 Distribution team load pressure prediction method and device, electronic equipment and storage medium
CN109698836A (en) * 2019-02-01 2019-04-30 重庆邮电大学 A kind of method for wireless lan intrusion detection and system based on deep learning
CN109738939A (en) * 2019-03-21 2019-05-10 蔡寅 A kind of Precursory Observational Data method for detecting abnormality
CN109740760A (en) * 2018-12-25 2019-05-10 平安科技（深圳）有限公司 Text quality inspection automates training method, electronic device and computer equipment
CN109754175A (en) * 2018-12-28 2019-05-14 广州明动软件股份有限公司 A Calculation Model and Its Application for Compressing and Predicting the Closing Time Limit of Administrative Examination and Approval Matters
CN109828888A (en) * 2019-01-28 2019-05-31 中国联合网络通信集团有限公司 Operation system method for monitoring state, device and computer readable storage medium
CN109978491A (en) * 2019-02-12 2019-07-05 平安科技（深圳）有限公司 Remind prediction technique, device, computer equipment and storage medium
CN109976908A (en) * 2019-03-15 2019-07-05 北京工业大学 A kind of server cluster dynamic retractility method based on RNN time series forecasting
CN110008079A (en) * 2018-12-25 2019-07-12 阿里巴巴集团控股有限公司 Monitor control index method for detecting abnormality, model training method, device and equipment
CN110148025A (en) * 2019-05-22 2019-08-20 郑州智通互联电子有限公司 A kind of scenic spot intelligence ticket sale system based on big data
CN110222840A (en) * 2019-05-17 2019-09-10 中山大学 A kind of cluster resource prediction technique and device based on attention mechanism
WO2019192136A1 (en) * 2018-04-03 2019-10-10 平安科技（深圳）有限公司 Electronic device, financial data processing method and system, and computer-readable storage medium
CN110472192A (en) * 2019-07-23 2019-11-19 平安科技（深圳）有限公司 Evaluation index extracting method, device, storage medium and server
CN110619423A (en) * 2019-08-06 2019-12-27 平安科技（深圳）有限公司 Multitask prediction method and device, electronic equipment and storage medium
CN110709861A (en) * 2018-03-13 2020-01-17 北京嘀嘀无限科技发展有限公司 Method and system for training a non-linear model
CN110808988A (en) * 2019-11-08 2020-02-18 国家计算机网络与信息安全管理中心山西分中心 Internet of things card service anomaly detection method based on information characteristic entropy and long-short term memory network
WO2020083381A1 (en) * 2018-10-25 2020-04-30 杭州海康威视数字技术股份有限公司 Database table area segmentation method and apparatus, device, and storage medium
CN111162949A (en) * 2019-12-31 2020-05-15 国网山西省电力公司信息通信分公司 An Interface Monitoring Method Based on Java Bytecode Embedding Technology
CN111221896A (en) * 2018-11-27 2020-06-02 北京京东尚科信息技术有限公司 User behavior prediction method and device, electronic device, storage medium
CN111241688A (en) * 2020-01-15 2020-06-05 北京百度网讯科技有限公司 Method and device for monitoring composite production process
CN111245667A (en) * 2018-11-28 2020-06-05 中国移动通信集团浙江有限公司 Network service identification method and device
CN111258854A (en) * 2020-01-21 2020-06-09 北京奇艺世纪科技有限公司 Model training method, alarm method based on prediction model and related device
WO2020133960A1 (en) * 2018-12-25 2020-07-02 平安科技（深圳）有限公司 Text quality inspection method, electronic apparatus, computer device and storage medium
CN111476403A (en) * 2020-03-17 2020-07-31 华为技术有限公司 Prediction model construction method and related device
CN111539479A (en) * 2020-04-27 2020-08-14 北京百度网讯科技有限公司 Method and apparatus for generating sample data
CN111652279A (en) * 2020-04-30 2020-09-11 中国平安财产保险股份有限公司 Behavior evaluation method and device based on time sequence data and readable storage medium
CN111666191A (en) * 2020-06-09 2020-09-15 贝壳技术有限公司 Data quality monitoring method and device, electronic equipment and storage medium
CN111737921A (en) * 2020-06-24 2020-10-02 深圳前海微众银行股份有限公司 Data processing method, equipment and medium based on recurrent neural network
CN111738422A (en) * 2020-06-24 2020-10-02 深圳前海微众银行股份有限公司 Data processing method, equipment and medium based on recurrent neural network
CN111858385A (en) * 2020-08-04 2020-10-30 深圳市汉云科技有限公司 Test method, device, equipment and storage medium of SQL database
CN112288126A (en) * 2020-09-09 2021-01-29 广东石油化工学院 Sampling data abnormal change online monitoring and diagnosing method
CN112419076A (en) * 2020-11-27 2021-02-26 好人生（上海）健康科技有限公司 Health insurance underwriting system and method based on big data and merchant insurance cloud platform
CN112506134A (en) * 2019-09-16 2021-03-16 阿里巴巴集团控股有限公司 Method, device and equipment for determining control variable value
CN112613995A (en) * 2020-12-30 2021-04-06 中国工商银行股份有限公司 Abnormality diagnosis method and apparatus
CN112785057A (en) * 2021-01-21 2021-05-11 上海东普信息科技有限公司 Component prediction method, device, equipment and storage medium based on exponential smoothing
CN113239272A (en) * 2021-05-12 2021-08-10 烽火通信科技股份有限公司 Intention prediction method and intention prediction device of network management and control system
CN113420902A (en) * 2020-12-09 2021-09-21 上海东普信息科技有限公司 Component prediction model training method, component prediction method and related equipment
CN113449923A (en) * 2021-07-09 2021-09-28 中国银行股份有限公司 Multi-model object market quotation prediction method and device
CN113505532A (en) * 2021-07-06 2021-10-15 新智数字科技有限公司 Method and device for predicting residual life of equipment, computer equipment and medium
CN113657628A (en) * 2021-08-20 2021-11-16 武汉霖汐科技有限公司 Industrial equipment monitoring method and system, electronic equipment and storage medium
CN113724795A (en) * 2020-05-25 2021-11-30 中国石油化工股份有限公司 Method and device for calculating concentration of raffinate hydrogen peroxide, storage medium and processor
CN113746575A (en) * 2021-09-03 2021-12-03 北京航空航天大学 Channel fading determination method and system for geostationary orbit satellite
CN114091930A (en) * 2021-11-25 2022-02-25 深圳前海微众银行股份有限公司 Business indicator early warning method, device, electronic device and storage medium
CN114266300A (en) * 2021-12-16 2022-04-01 中国联合网络通信集团有限公司 Feature prediction model training, core network service anomaly detection method and device
CN114358353A (en) * 2020-09-27 2022-04-15 中国移动通信集团浙江有限公司 Method and device for predicting complaining worker singularity and computing equipment
CN114445143A (en) * 2022-01-29 2022-05-06 中国农业银行股份有限公司 A business data prediction method, device, equipment and medium
CN114581159A (en) * 2022-05-04 2022-06-03 爱迪森(北京)生物科技有限公司 Warehouse prediction method and system based on big data analysis and readable storage medium
CN115470936A (en) * 2022-09-23 2022-12-13 广州爱浦路网络技术有限公司 NWDAF-based machine learning model updating method and device
CN115600728A (en) * 2022-09-23 2023-01-13 广东邦普循环科技有限公司（Cn） Method and device for estimating annual carbon emissions of power batteries
WO2023283953A1 (en) * 2021-07-16 2023-01-19 北京小米移动软件有限公司 Communication processing method and apparatus, communication device, and storage medium
CN116703470A (en) * 2023-08-09 2023-09-05 深圳市土地公网络科技有限公司 Method, device, equipment and storage medium for predicting supply information
CN117289671A (en) * 2023-11-27 2023-12-26 博纯材料股份有限公司 State monitoring method and system of Gao Jiezhe alkane purification production control system

Families Citing this family (87)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110163471B (en) * 2019-04-10 2023-08-18 创新先进技术有限公司 Abnormality identification method and device
CN110288193B (en) * 2019-05-23 2024-04-09 中国平安人寿保险股份有限公司 Task monitoring processing method and device, computer equipment and storage medium
CN110188940B (en) * 2019-05-24 2023-09-05 深圳市腾讯计算机系统有限公司 Service processing method and device
CN112148765B (en) * 2019-06-28 2024-04-09 北京百度网讯科技有限公司 Service data processing method, device and storage medium
CN110309153B (en) * 2019-06-28 2023-07-14 携程旅游信息技术（上海）有限公司 Method, system, equipment and storage medium for processing monitoring data of call center
CN112306788A (en) * 2019-07-29 2021-02-02 北京京东振世信息技术有限公司 Program dynamic monitoring method and device
CN112435035B (en) * 2019-08-09 2025-07-08 阿里巴巴集团控股有限公司 Data auditing method, device and equipment
CN110705598B (en) * 2019-09-06 2024-05-28 中国平安财产保险股份有限公司 Intelligent model management method, intelligent model management device, computer equipment and storage medium
CN112529236B (en) * 2019-09-18 2024-10-25 泰康保险集团股份有限公司 Target object identification method, device, electronic equipment and storage medium
CN112686418B (en) * 2019-10-18 2024-07-16 北京京东振世信息技术有限公司 Method and device for predicting performance aging
CN110866528B (en) * 2019-10-28 2023-11-28 腾讯科技（深圳）有限公司 A model training method, energy consumption efficiency prediction method, device and medium
CN112835780B (en) * 2019-11-25 2024-02-02 杭州海康威视系统技术有限公司 Service detection method and device
CN111026626A (en) * 2019-11-29 2020-04-17 中国建设银行股份有限公司 CPU consumption estimation and estimation model training method and device
CN112884159B (en) * 2019-11-30 2024-06-18 华为技术有限公司 Model updating system, model updating method and related equipment
CN112990520B (en) * 2019-12-13 2024-08-20 顺丰科技有限公司 Method, device, computer equipment and storage medium for predicting net point connection piece quantity
CN111178568B (en) * 2019-12-31 2023-06-23 中国银行股份有限公司 Service reservation request processing method, device, server and storage medium
CN113313439B (en) * 2020-02-26 2024-04-05 北京京东振世信息技术有限公司 Method and device for calculating time length of tall-in-hand
EP4113483A4 (en) * 2020-02-28 2024-03-13 Daikin Industries, Ltd. EFFICIENCY ESTIMATION DEVICE
CN111460913B (en) * 2020-03-13 2024-12-24 北京理工大学 An adaptive directional forecasting method for weather data
CN111461862B (en) * 2020-03-27 2023-06-30 支付宝(杭州)信息技术有限公司 Method and device for determining target characteristics for service data
CN111767938B (en) * 2020-05-09 2023-12-19 北京奇艺世纪科技有限公司 Abnormal data detection method and device and electronic equipment
CN113743688B (en) * 2020-05-27 2023-10-20 富联精密电子(天津)有限公司 Quality control method, quality control device, computer device and storage medium
CN111694814B (en) * 2020-05-27 2024-04-09 平安银行股份有限公司 Batch expansion method and device for date partition table, computer equipment and storage medium
CN111708561B (en) * 2020-06-17 2024-01-05 杭州海康消防科技有限公司 Algorithm model updating system, method and device and electronic equipment
CN111738463A (en) * 2020-06-17 2020-10-02 深圳华远云联数据科技有限公司 Operation and maintenance method, device, system, electronic equipment and storage medium
CN111859783B (en) * 2020-06-22 2023-10-13 重庆大学 Water pressure prediction methods, systems, storage media, equipment, urban water supply systems
CN111723942B (en) * 2020-06-29 2024-02-02 南方电网科学研究院有限责任公司 Enterprise electricity load prediction method, power grid business subsystem and prediction system
CN111783040B (en) * 2020-06-30 2025-01-28 深圳前海微众银行股份有限公司 A method and device for testing and evaluating service performance stability
CN111756646B (en) * 2020-07-08 2023-09-29 腾讯科技（深圳）有限公司 Network transmission control method, device, computer equipment and storage medium
CN113936331A (en) * 2020-07-14 2022-01-14 香港理工大学深圳研究院 Behavior recognition method and device, terminal equipment and storage medium
CN111833114A (en) * 2020-07-27 2020-10-27 北京思特奇信息技术股份有限公司 A method, system, medium and device for intelligent prediction of channel business development target
CN111985713B (en) * 2020-08-19 2023-08-18 中国银行股份有限公司 Data index waveform prediction method and device
CN113297542B (en) * 2020-08-20 2025-01-14 湖南长天自控工程有限公司 A method and device for predicting raw ball ratio of a ball making machine
CN112132394B (en) * 2020-08-21 2024-03-29 西安交通大学 Power plant circulating water pump predictive state evaluation method and system
CN114118502B (en) * 2020-08-26 2025-08-08 顺丰科技有限公司 Customer call prediction method, device, computer equipment and storage medium
CN112288172A (en) * 2020-10-30 2021-01-29 国家电网有限公司 Method and device for predicting line loss rate in station area
CN112613224A (en) * 2020-11-25 2021-04-06 西人马联合测控（泉州）科技有限公司 Training method, detection method, device and equipment of bridge structure detection model
CN112529158B (en) * 2020-12-08 2025-02-21 华强方特（深圳）科技有限公司 Position prediction method, device, computer equipment and storage medium
CN112560085B (en) * 2020-12-10 2023-09-19 支付宝(杭州)信息技术有限公司 Privacy protection method and device for business prediction model
CN112529624B (en) * 2020-12-15 2024-01-09 北京百度网讯科技有限公司 Method, device, equipment and storage medium for generating business prediction model
CN112560988B (en) * 2020-12-25 2023-09-19 竹间智能科技（上海）有限公司 Model training method and device
CN112650661B (en) * 2020-12-29 2024-07-09 北京嘀嘀无限科技发展有限公司 Data processing quality control method, device, computer equipment and storage medium
US12322111B2 (en) 2020-12-30 2025-06-03 United Imaging Research Institute of Innovative Medical Equipment Image segmentation method, device, equipment and storage medium
CN112801940B (en) * 2020-12-31 2024-07-02 深圳市联影高端医疗装备创新研究院 Model evaluation method, device, equipment and medium
CN113762688B (en) * 2021-01-06 2024-12-03 北京沃东天骏信息技术有限公司 Business analysis system, method and storage medium
CN114764741A (en) * 2021-01-15 2022-07-19 深圳光耀智微科技有限公司 Method and system for predicting running wind power of wind driven generator
CN113743642B (en) * 2021-01-27 2024-08-20 北京沃东天骏信息技术有限公司 Predictive model training method and device and touch number prediction method and device
CN112766396A (en) * 2021-01-27 2021-05-07 昆仑数智科技有限责任公司 System, method, computer device and medium for detecting device abnormality
CN112758100B (en) * 2021-02-03 2023-03-14 洪丰 Accelerator mistaken stepping detection method and device
CN112819239B (en) * 2021-02-19 2024-05-14 北京骑胜科技有限公司 Product quantity prediction method, device, electronic equipment and readable storage medium
CN113704082B (en) * 2021-02-26 2025-02-07 腾讯科技（深圳）有限公司 Model evaluation method, device, electronic device and storage medium
CN113256328B (en) * 2021-05-18 2024-02-23 深圳索信达数据技术有限公司 Method, device, computer equipment and storage medium for predicting target clients
CN113222057A (en) * 2021-05-28 2021-08-06 中邮信息科技(北京)有限公司 Data prediction model training method, data prediction device, data prediction equipment and data prediction medium
CN115484188A (en) * 2021-06-16 2022-12-16 中国移动通信集团广东有限公司 TAP device monitoring method and system, electronic device and readable storage medium
CN113379139A (en) * 2021-06-22 2021-09-10 阳光电源股份有限公司 Charging station management method and application device thereof
CN113486584B (en) * 2021-07-06 2023-12-29 新奥新智科技有限公司 Method and device for predicting equipment failure, computer equipment and computer readable storage medium
CN113626508B (en) * 2021-07-13 2024-08-06 交控科技股份有限公司 Train feature library management method and device, electronic equipment and readable storage medium
CN113673080B (en) * 2021-07-15 2023-08-18 北京奇艺世纪科技有限公司 Method, device, equipment and medium for predicting duration of application and training model
CN115700953A (en) * 2021-07-23 2023-02-07 中国科学院微电子研究所 Energy control method and device for excimer laser, computer equipment and medium
CN113837527A (en) * 2021-08-02 2021-12-24 深圳前海微众银行股份有限公司 Enterprise rating method, apparatus, equipment, and storage medium
CN113792906B (en) * 2021-08-05 2024-04-30 交控科技股份有限公司 Train long-time window running track prediction method, device, equipment and storage medium
CN113506175A (en) * 2021-08-19 2021-10-15 北京中数智汇科技股份有限公司 Method, device, equipment and storage medium for optimizing risk early warning model of medium and small enterprises
CN113673866A (en) * 2021-08-20 2021-11-19 上海寻梦信息技术有限公司 Crop decision method, model training method and related equipment
CN113703923B (en) * 2021-08-31 2024-05-28 深信服科技股份有限公司 Service problem identification method, device, equipment and medium
CN113988369B (en) * 2021-09-23 2025-04-11 上海三一重机股份有限公司 Training method of prediction model and prediction method based on prediction model
CN113822580B (en) * 2021-09-24 2024-06-28 深圳市出新知识产权管理有限公司 Equipment working condition assessment method and related equipment
CN113837289B (en) * 2021-09-26 2024-03-19 创新奇智(重庆)科技有限公司 Model training method, fault prediction device and electronic equipment
CN114118445B (en) * 2021-12-02 2024-09-20 深圳市华傲数据技术有限公司 Time sequence data processing method, equipment and medium oriented to machine learning modeling
CN114219562A (en) * 2021-12-13 2022-03-22 香港中文大学（深圳） Model training method, enterprise credit evaluation method and device, equipment and medium
CN115002513B (en) * 2022-05-25 2023-10-20 咪咕文化科技有限公司 Audio and video scheduling method and device, electronic equipment and computer readable storage medium
CN117273182B (en) * 2022-06-08 2025-02-14 腾讯科技（深圳）有限公司 Parameter prediction method, prediction server, prediction system and electronic equipment
CN115169605B (en) * 2022-06-28 2023-06-20 国网山东省电力公司兰陵县供电公司 Monitoring method and system for substation primary equipment
CN115372200A (en) * 2022-07-12 2022-11-22 蓝冰河(常州)精密测量技术有限责任公司 On-line surface density measurement method and device based on X-ray
CN115426044B (en) * 2022-07-12 2025-08-12 北京邮电大学 Compensation method for optical fiber frequency transmission and related equipment
CN115391208A (en) * 2022-08-30 2022-11-25 中国银行股份有限公司 A test data acquisition method, device, equipment and storage medium
CN115659881A (en) * 2022-10-18 2023-01-31 中国科学院微电子研究所 Method for acquiring circuit unit characteristic data
CN115600747B (en) * 2022-10-24 2024-02-13 无锡瑞鼎电力科技有限公司 Tunnel state monitoring and management method and system based on Internet of things
CN115630830A (en) * 2022-12-01 2023-01-20 北京忠业兴达科技有限公司 Power supply and distribution method, device, equipment and storage medium for data center
CN116016223B (en) * 2022-12-09 2024-02-02 国网湖北省电力有限公司信息通信公司 Data transmission optimization method for data center network
CN116702072A (en) * 2023-05-16 2023-09-05 中国联合网络通信集团有限公司 Method and device for determining abnormal data and storage medium
CN116500426B (en) * 2023-06-28 2023-09-05 东莞市兆恒机械有限公司 Method for calibrating high-temperature test of semiconductor detection equipment
CN116865787B (en) * 2023-07-06 2024-02-09 北京煜邦电力技术股份有限公司 Method for intelligently selecting transmission opportunity of power carrier signal
CN116627991B (en) * 2023-07-26 2023-09-26 山东朝阳轴承有限公司 Enterprise informatization data storage method and system based on Internet of things
CN117370272B (en) * 2023-10-25 2024-06-11 浙江星汉信息技术股份有限公司 File management method, device, equipment and storage medium based on file heat
CN117335416B (en) * 2023-11-24 2024-03-01 国网浙江省电力有限公司 Method, device, equipment and storage medium for optimizing power load
CN117932280B (en) * 2024-03-25 2024-06-25 之江实验室 Long sequence data prediction method, long sequence data prediction device, computer equipment, medium and long sequence data prediction product
CN118735496B (en) * 2024-06-06 2025-02-18 浙江优财云链科技有限公司 Front-end processor data system and method of bank system based on parameter driving

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150088606A1 (en) * 2013-09-20 2015-03-26 Tata Consultancy Services Ltd. Computer Implemented Tool and Method for Automating the Forecasting Process
CN105786830A (en) * 2014-12-19 2016-07-20 阿里巴巴集团控股有限公司 Method, device and system for self-adaptively adjusting models in computer systems
CN106603293A (en) * 2016-12-20 2017-04-26 南京邮电大学 Network fault diagnosis method based on deep learning in virtual network environment
CN106886846A (en) * 2017-04-26 2017-06-23 中南大学 A kind of bank outlets' excess reserve Forecasting Methodology that Recognition with Recurrent Neural Network is remembered based on shot and long term

Cited By (103)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110709861A (en) * 2018-03-13 2020-01-17 北京嘀嘀无限科技发展有限公司 Method and system for training a non-linear model
CN108549954A (en) * 2018-03-26 2018-09-18 平安科技（深圳）有限公司 Risk model training method, risk identification method, device, equipment and medium
CN108549954B (en) * 2018-03-26 2022-08-02 平安科技（深圳）有限公司 Risk model training method, risk identification device, risk identification equipment and risk identification medium
CN108427854A (en) * 2018-03-29 2018-08-21 深圳市智物联网络有限公司 A kind of data analysing method and its relevant device
CN108427854B (en) * 2018-03-29 2022-06-07 深圳市智物联网络有限公司 Data analysis method and related equipment thereof
CN108446771A (en) * 2018-04-02 2018-08-24 四川长虹电器股份有限公司 A method of preventing Sale Forecasting Model over-fitting
CN108830723A (en) * 2018-04-03 2018-11-16 平安科技（深圳）有限公司 Electronic device, bond yield analysis method and storage medium
WO2019192136A1 (en) * 2018-04-03 2019-10-10 平安科技（深圳）有限公司 Electronic device, financial data processing method and system, and computer-readable storage medium
CN108446864B (en) * 2018-04-10 2022-03-29 广州新科佳都科技有限公司 Big data analysis-based fault early warning system and method for rail transit equipment
CN108446864A (en) * 2018-04-10 2018-08-24 广州新科佳都科技有限公司 The fault early warning system and method for Transit Equipment based on big data analysis
CN108615360A (en) * 2018-05-08 2018-10-02 东南大学 Transport need based on neural network Evolution Forecast method day by day
CN108615360B (en) * 2018-05-08 2022-02-11 东南大学 Prediction method of daily evolution of traffic demand based on neural network
CN108615096A (en) * 2018-05-10 2018-10-02 平安科技（深圳）有限公司 Server, the processing method of Financial Time Series and storage medium
CN108764974A (en) * 2018-05-11 2018-11-06 国网电子商务有限公司 A kind of procurement of commodities amount prediction technique and device based on deep learning
CN108764572A (en) * 2018-05-28 2018-11-06 安徽磐众信息科技有限公司 A kind of finance data prediction technique based on Recognition with Recurrent Neural Network
CN108846692A (en) * 2018-06-05 2018-11-20 浙江大学城市学院 A kind of consumer spending behavior prediction method based on multifactor Recognition with Recurrent Neural Network
CN108846692B (en) * 2018-06-05 2021-03-30 浙江大学城市学院 A prediction method of consumer consumption behavior based on multi-factor recurrent neural network
CN109028480A (en) * 2018-06-08 2018-12-18 杭州古北电子科技有限公司 A kind of thermostatic constant wet control system and its method
CN109028480B (en) * 2018-06-08 2021-08-20 杭州博联智能科技股份有限公司 Constant temperature and humidity control system and method thereof
CN108977854A (en) * 2018-08-07 2018-12-11 中铝视拓智能科技有限公司 Electrolyzer temperature monitoring method, device, equipment, system and readable storage medium storing program for executing
CN109272344A (en) * 2018-08-07 2019-01-25 阿里巴巴集团控股有限公司 Model training method and device, data predication method and device, server
CN109194423A (en) * 2018-08-13 2019-01-11 中国人民解放军陆军工程大学 Single-frequency point spectrum prediction method based on optimized long-short term memory model
CN109190808B (en) * 2018-08-15 2020-11-24 拍拍信数据服务（上海）有限公司 User behavior prediction method, device, equipment and medium
CN109190808A (en) * 2018-08-15 2019-01-11 拍拍信数据服务（上海）有限公司 User's behavior prediction method, apparatus, equipment and medium
CN109308226A (en) * 2018-08-22 2019-02-05 中国平安人寿保险股份有限公司 Data exception determines method and device, storage medium and electronic equipment
CN109034500A (en) * 2018-09-04 2018-12-18 湘潭大学 A kind of mid-term electric load forecasting method of multiple timings collaboration
CN109345373A (en) * 2018-09-11 2019-02-15 北京三快在线科技有限公司 Write-off risk early warning method, device, electronic device and computer-readable medium
CN109660419A (en) * 2018-10-08 2019-04-19 平安科技（深圳）有限公司 Predict method, apparatus, equipment and the storage medium of network equipment exception
WO2020083381A1 (en) * 2018-10-25 2020-04-30 杭州海康威视数字技术股份有限公司 Database table area segmentation method and apparatus, device, and storage medium
CN111104569B (en) * 2018-10-25 2023-10-20 杭州海康威视数字技术股份有限公司 Method, device and storage medium for partitioning database table
CN111104569A (en) * 2018-10-25 2020-05-05 杭州海康威视数字技术股份有限公司 Region segmentation method and device for database table and storage medium
CN109378065A (en) * 2018-10-30 2019-02-22 医渡云（北京）技术有限公司 Medical data processing method and processing device, storage medium, electronic equipment
CN109672795B (en) * 2018-11-14 2022-03-11 平安科技（深圳）有限公司 Call center resource management method and device, electronic equipment and storage medium
CN109672795A (en) * 2018-11-14 2019-04-23 平安科技（深圳）有限公司 Call center resource management method and device, electronic equipment, storage medium
CN111221896B (en) * 2018-11-27 2025-01-17 北京京东尚科信息技术有限公司 User behavior prediction method and device, electronic equipment and storage medium
CN111221896A (en) * 2018-11-27 2020-06-02 北京京东尚科信息技术有限公司 User behavior prediction method and device, electronic device, storage medium
CN111245667A (en) * 2018-11-28 2020-06-05 中国移动通信集团浙江有限公司 Network service identification method and device
CN109740760B (en) * 2018-12-25 2024-04-05 平安科技（深圳）有限公司 Text quality inspection automatic training method, electronic device and computer equipment
CN110008079A (en) * 2018-12-25 2019-07-12 阿里巴巴集团控股有限公司 Monitor control index method for detecting abnormality, model training method, device and equipment
CN109740760A (en) * 2018-12-25 2019-05-10 平安科技（深圳）有限公司 Text quality inspection automates training method, electronic device and computer equipment
WO2020133960A1 (en) * 2018-12-25 2020-07-02 平安科技（深圳）有限公司 Text quality inspection method, electronic apparatus, computer device and storage medium
CN109685275A (en) * 2018-12-27 2019-04-26 拉扎斯网络科技（上海）有限公司 Distribution team load pressure prediction method and device, electronic equipment and storage medium
CN109754175A (en) * 2018-12-28 2019-05-14 广州明动软件股份有限公司 A Calculation Model and Its Application for Compressing and Predicting the Closing Time Limit of Administrative Examination and Approval Matters
CN109828888A (en) * 2019-01-28 2019-05-31 中国联合网络通信集团有限公司 Operation system method for monitoring state, device and computer readable storage medium
CN109698836A (en) * 2019-02-01 2019-04-30 重庆邮电大学 A kind of method for wireless lan intrusion detection and system based on deep learning
CN109978491B (en) * 2019-02-12 2024-02-06 平安科技（深圳）有限公司 Reminding prediction method, reminding prediction device, computer equipment and storage medium
CN109978491A (en) * 2019-02-12 2019-07-05 平安科技（深圳）有限公司 Remind prediction technique, device, computer equipment and storage medium
CN109976908A (en) * 2019-03-15 2019-07-05 北京工业大学 A kind of server cluster dynamic retractility method based on RNN time series forecasting
CN109738939A (en) * 2019-03-21 2019-05-10 蔡寅 A kind of Precursory Observational Data method for detecting abnormality
CN110222840A (en) * 2019-05-17 2019-09-10 中山大学 A kind of cluster resource prediction technique and device based on attention mechanism
CN110222840B (en) * 2019-05-17 2023-05-05 中山大学 A method and device for cluster resource prediction based on attention mechanism
CN110148025A (en) * 2019-05-22 2019-08-20 郑州智通互联电子有限公司 A kind of scenic spot intelligence ticket sale system based on big data
CN110472192B (en) * 2019-07-23 2022-04-15 平安科技（深圳）有限公司 Evaluation index extraction method, evaluation index extraction device, storage medium, and server
CN110472192A (en) * 2019-07-23 2019-11-19 平安科技（深圳）有限公司 Evaluation index extracting method, device, storage medium and server
CN110619423A (en) * 2019-08-06 2019-12-27 平安科技（深圳）有限公司 Multitask prediction method and device, electronic equipment and storage medium
CN112506134A (en) * 2019-09-16 2021-03-16 阿里巴巴集团控股有限公司 Method, device and equipment for determining control variable value
CN110808988A (en) * 2019-11-08 2020-02-18 国家计算机网络与信息安全管理中心山西分中心 Internet of things card service anomaly detection method based on information characteristic entropy and long-short term memory network
CN110808988B (en) * 2019-11-08 2021-09-10 国家计算机网络与信息安全管理中心山西分中心 Internet of things card service anomaly detection method based on information characteristic entropy and long-short term memory network
CN111162949A (en) * 2019-12-31 2020-05-15 国网山西省电力公司信息通信分公司 An Interface Monitoring Method Based on Java Bytecode Embedding Technology
CN111241688B (en) * 2020-01-15 2023-08-25 北京百度网讯科技有限公司 Method and device for monitoring composite production process
CN111241688A (en) * 2020-01-15 2020-06-05 北京百度网讯科技有限公司 Method and device for monitoring composite production process
CN111258854B (en) * 2020-01-21 2023-10-03 北京奇艺世纪科技有限公司 Model training method, alarm method based on prediction model and related device
CN111258854A (en) * 2020-01-21 2020-06-09 北京奇艺世纪科技有限公司 Model training method, alarm method based on prediction model and related device
CN111476403A (en) * 2020-03-17 2020-07-31 华为技术有限公司 Prediction model construction method and related device
CN111539479B (en) * 2020-04-27 2023-08-08 北京百度网讯科技有限公司 Method and apparatus for generating sample data
CN111539479A (en) * 2020-04-27 2020-08-14 北京百度网讯科技有限公司 Method and apparatus for generating sample data
CN111652279A (en) * 2020-04-30 2020-09-11 中国平安财产保险股份有限公司 Behavior evaluation method and device based on time sequence data and readable storage medium
CN111652279B (en) * 2020-04-30 2024-04-30 中国平安财产保险股份有限公司 Behavior evaluation method and device based on time sequence data and readable storage medium
CN113724795A (en) * 2020-05-25 2021-11-30 中国石油化工股份有限公司 Method and device for calculating concentration of raffinate hydrogen peroxide, storage medium and processor
CN111666191B (en) * 2020-06-09 2023-09-29 贝壳技术有限公司 Data quality monitoring method and device, electronic equipment and storage medium
CN111666191A (en) * 2020-06-09 2020-09-15 贝壳技术有限公司 Data quality monitoring method and device, electronic equipment and storage medium
CN111738422A (en) * 2020-06-24 2020-10-02 深圳前海微众银行股份有限公司 Data processing method, equipment and medium based on recurrent neural network
CN111737921A (en) * 2020-06-24 2020-10-02 深圳前海微众银行股份有限公司 Data processing method, equipment and medium based on recurrent neural network
CN111737921B (en) * 2020-06-24 2024-04-26 深圳前海微众银行股份有限公司 Data processing method, equipment and medium based on cyclic neural network
CN111858385A (en) * 2020-08-04 2020-10-30 深圳市汉云科技有限公司 Test method, device, equipment and storage medium of SQL database
CN111858385B (en) * 2020-08-04 2024-02-20 深圳市汉云科技有限公司 SQL database testing method, device, equipment and storage medium
CN112288126A (en) * 2020-09-09 2021-01-29 广东石油化工学院 Sampling data abnormal change online monitoring and diagnosing method
CN114358353A (en) * 2020-09-27 2022-04-15 中国移动通信集团浙江有限公司 Method and device for predicting complaining worker singularity and computing equipment
CN112419076A (en) * 2020-11-27 2021-02-26 好人生（上海）健康科技有限公司 Health insurance underwriting system and method based on big data and merchant insurance cloud platform
CN113420902A (en) * 2020-12-09 2021-09-21 上海东普信息科技有限公司 Component prediction model training method, component prediction method and related equipment
CN112613995A (en) * 2020-12-30 2021-04-06 中国工商银行股份有限公司 Abnormality diagnosis method and apparatus
CN112785057A (en) * 2021-01-21 2021-05-11 上海东普信息科技有限公司 Component prediction method, device, equipment and storage medium based on exponential smoothing
CN112785057B (en) * 2021-01-21 2023-09-01 上海东普信息科技有限公司 Method, device, equipment and storage medium for predicting quantity of parts based on exponential smoothing
CN113239272A (en) * 2021-05-12 2021-08-10 烽火通信科技股份有限公司 Intention prediction method and intention prediction device of network management and control system
CN113505532B (en) * 2021-07-06 2023-12-29 新奥新智科技有限公司 Equipment remaining life prediction method, device, computer equipment and medium
CN113505532A (en) * 2021-07-06 2021-10-15 新智数字科技有限公司 Method and device for predicting residual life of equipment, computer equipment and medium
CN113449923A (en) * 2021-07-09 2021-09-28 中国银行股份有限公司 Multi-model object market quotation prediction method and device
WO2023283953A1 (en) * 2021-07-16 2023-01-19 北京小米移动软件有限公司 Communication processing method and apparatus, communication device, and storage medium
CN113657628A (en) * 2021-08-20 2021-11-16 武汉霖汐科技有限公司 Industrial equipment monitoring method and system, electronic equipment and storage medium
CN113746575A (en) * 2021-09-03 2021-12-03 北京航空航天大学 Channel fading determination method and system for geostationary orbit satellite
CN114091930B (en) * 2021-11-25 2024-06-18 深圳前海微众银行股份有限公司 Service index early warning method and device, electronic equipment and storage medium
CN114091930A (en) * 2021-11-25 2022-02-25 深圳前海微众银行股份有限公司 Business indicator early warning method, device, electronic device and storage medium
CN114266300A (en) * 2021-12-16 2022-04-01 中国联合网络通信集团有限公司 Feature prediction model training, core network service anomaly detection method and device
CN114266300B (en) * 2021-12-16 2023-01-24 中国联合网络通信集团有限公司 Feature prediction model training, core network service anomaly detection method and device
CN114445143A (en) * 2022-01-29 2022-05-06 中国农业银行股份有限公司 A business data prediction method, device, equipment and medium
CN114581159B (en) * 2022-05-04 2022-08-12 爱迪森(北京)生物科技有限公司 Warehouse prediction method and system based on big data analysis and readable storage medium
CN114581159A (en) * 2022-05-04 2022-06-03 爱迪森(北京)生物科技有限公司 Warehouse prediction method and system based on big data analysis and readable storage medium
CN115600728A (en) * 2022-09-23 2023-01-13 广东邦普循环科技有限公司（Cn） Method and device for estimating annual carbon emissions of power batteries
CN115470936A (en) * 2022-09-23 2022-12-13 广州爱浦路网络技术有限公司 NWDAF-based machine learning model updating method and device
CN116703470A (en) * 2023-08-09 2023-09-05 深圳市土地公网络科技有限公司 Method, device, equipment and storage medium for predicting supply information
CN116703470B (en) * 2023-08-09 2024-01-12 深圳市土地公网络科技有限公司 Method, device, equipment and storage medium for predicting supply information
CN117289671A (en) * 2023-11-27 2023-12-26 博纯材料股份有限公司 State monitoring method and system of Gao Jiezhe alkane purification production control system
CN117289671B (en) * 2023-11-27 2024-02-02 博纯材料股份有限公司 State monitoring method and system of Gao Jiezhe alkane purification production control system

Also Published As

Publication number Publication date
WO2019056499A1 (en) 2019-03-28

Similar Documents

Publication Publication Date Title
CN107730087A (en) Forecast model training method, data monitoring method, device, equipment and medium
Tang et al. ST‐LSTM: A Deep Learning Approach Combined Spatio‐Temporal Features for Short‐Term Forecast in Rail Transit
US10387900B2 (en) Methods and apparatus for self-adaptive time series forecasting engine
Chen et al. Bayesian forecasting for financial risk management, pre and post the global financial crisis
AU2022211812A1 (en) Method and system of dynamic model selection for time series forecasting
Zhang et al. Comparison of econometric models and artificial neural networks algorithms for the prediction of baltic dry index
CN109726865A (en) User load probability density prediction method, device and storage medium based on EMD-QRF
CN110852881A (en) Risk account identification method and device, electronic equipment and medium
Vu et al. Short-term forecasting of electricity spot prices containing random spikes using a time-varying autoregressive model combined with kernel regression
Lin et al. Tourism demand forecasting: Econometric model based on multivariate adaptive regression splines, artificial neural network and support vector regression
Kakade et al. Forecasting commodity market returns volatility: a hybrid ensemble learning garch‐lstm based approach
CN112561320A (en) Training method of mechanism risk prediction model, mechanism risk prediction method and device
Wijesinghe Time series forecasting: Analysis of LSTM neural networks to predict exchange rates of currencies
CN113034284A (en) Stock tendency analysis and early warning system based on algorithm, big data and block chain
Selvakumar et al. Divination of stock market exploration using long short-term memory (LSTM)
CN110866672A (en) Data processing method, device, terminal and medium
US20160125306A1 (en) Goal-attainment assessment apparatus and method
CN114282657A (en) Market data long-term prediction model training method, device, equipment and storage medium
Shafiei et al. Integration of departure time choice modeling and dynamic origin–destination demand estimation in a large-scale network
CN114418776A (en) Data processing method, device, terminal equipment and medium
Tiwari et al. Modelling of ambient noise levels in urban environment
CN116703216A (en) Service scene scoring method, device, computer equipment and storage medium
CN116739649A (en) User response potential evaluation method and device
Krishna et al. Identifying demand forecasting using machine learning for business intelligence
CN112101611B (en) Real estate customer buyback time prediction method, server and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20180223

RJ01 Rejection of invention patent application after publication
点击 这是indexloc提供的php浏览器服务，不要输入任何密码和下载