CN110765330A - Data set classification method, apparatus, equipment and computer storage medium - Google Patents

Abstract

The invention relates to the technical field of financial technology (Fintech), and discloses a data set classification method. The method includes: pre-dividing a data set to be classified into a training set, a test set and a verification set, and based on the training set, all The test set and the verification set establish a dynamic image; the training scale value corresponding to the training set is adjusted according to the adjustment instruction received by the dynamic image to obtain the target training scale value; based on the target training scale value and The preset constraint conditions respectively update the test scale value corresponding to the test set and the verification scale value corresponding to the verification set, and based on the target training scale value, the updated test scale value and the updated test scale value. The validation scale value classifies the dataset. The invention also discloses a data set classification device, equipment and a computer storage medium. The invention improves the processing efficiency of classifying the data set.

Description

Data set classification method, apparatus, equipment and computer storage medium

technical field

The present invention relates to the technical field of financial technology (Fintech), and in particular, to a data set classification method, apparatus, device and computer storage medium.

Background technique

With the development of computer technology, more and more technologies (big data, distributed, blockchain, artificial intelligence, etc.) are applied in the financial field, and the traditional financial industry is gradually transforming into financial technology (Fintech). The security and real-time requirements of the industry also put forward higher requirements for technology. For example, in the process of machine learning, it is often necessary to classify a data set into a training set, a test set and a validation set for training separately, of which the training set is a necessary data set, and the test set and validation set can be divided according to actual purposes. . At present, the conventional method is to set the numerical ratio of 3 data sets to the total data set respectively, and the sum of the 3 ratios must be 100%. However, when adjusting the proportion of a certain data set in this method, it is necessary to manually adjust the other two The proportion of data sets is very inconvenient to operate, and the input error rate is high, resulting in low processing efficiency of data sets.

SUMMARY OF THE INVENTION

The main purpose of the present invention is to provide a data set classification method, device, equipment and computer storage medium, aiming at improving the processing efficiency of the data set.

In order to achieve the above object, the present invention provides a data set classification method, and the data set classification method includes the following steps:

Pre-dividing the data set to be classified into a training set, a test set and a verification set, and creating a dynamic image based on the training set, the test set and the verification set;

Adjust the training scale value corresponding to the training set according to the adjustment instruction received from the dynamic image to obtain the target training scale value;

The test scale value corresponding to the test set and the verification scale value corresponding to the verification set are respectively updated based on the target training scale value and preset constraints, and based on the target training scale value, the updated all scale values The data set is classified according to the test scale value and the updated validation scale value.

Optionally, the step of establishing a dynamic image based on the training set, the test set and the verification set includes:

Obtain the training ratio value of the training set in the data set, the test ratio value of the test set in the data set and the verification ratio value of the verification set in the data set;

The training scale value, the test scale value and the verification scale value are transferred into a preset image to obtain a dynamic image.

Optionally, the step of respectively updating the test scale value corresponding to the test set and the verification scale value corresponding to the verification set based on the target training scale value and a preset constraint condition includes:

Calculate the remaining ratio value of the data set according to the target training ratio value, and based on preset constraints and the remaining ratio value, respectively, the test ratio value corresponding to the test set and the verification ratio corresponding to the verification set value is updated.

Optionally, the constraint condition includes equal-scale adjustment, and the test scale value corresponding to the test set and the verification scale value corresponding to the verification set are respectively updated based on the preset constraint condition and the remaining scale value. steps, including:

Obtain the first sum value between the test scale value corresponding to the test set and the verification scale value corresponding to the verification set, and determine whether the first sum value is greater than the remaining scale value;

If it is less than, the test proportional value and the verification proportional value are simultaneously increased and updated in equal proportions based on a preset equal proportional adjustment, until the first sum value is equal to the remaining proportional value.

Optionally, after the step of judging whether the first sum value is greater than the remaining ratio value, the step includes:

If it is greater than the value, the test proportional value and the verification proportional value are simultaneously subjected to proportional reduction and update processing based on a preset equal proportional adjustment, until the first sum value is equal to the remaining proportional value.

Optionally, after the step of establishing a dynamic image based on the training set, the test set and the verification set, it includes:

Obtain the training scale value corresponding to the training set in the dynamic image, the test scale value corresponding to the test set and the verification scale value corresponding to the verification set;

receiving a test scale value update instruction, and updating the test scale value based on the test scale value update instruction to obtain an updated test scale value;

The verification scale value is updated according to the training scale value and the update test scale value to obtain a new verification scale value, and based on the training scale value, the update test scale value and the new verification scale value Classify the dataset.

Optionally, after the step of establishing a dynamic image based on the training set, the test set and the verification set, it includes:

Obtain the training scale value corresponding to the training set in the dynamic image, the test scale value corresponding to the test set and the verification scale value corresponding to the verification set;

receiving a verification scale value update instruction, and updating the verification scale value based on the verification scale value update instruction to obtain a new verification scale value;

The test scale value is updated based on the training scale value and the new validation scale value to obtain a new test scale value, and based on the training scale value, the new validation scale value and the new test The scale value classifies the dataset.

In addition, in order to achieve the above object, the present invention also provides a data set classification device, the data set classification device includes:

establishing a module for pre-dividing the data set to be classified into a training set, a test set and a verification set, and establishing a dynamic image based on the training set, the test set and the verification set;

an acquisition module, configured to adjust the training scale value corresponding to the training set according to the adjustment instruction received by the dynamic image, so as to obtain the target training scale value;

A classification module, configured to update the test scale value corresponding to the test set and the verification scale value corresponding to the verification set based on the target training scale value, and based on the target training scale value, the updated test scale value The scale value and the updated validation scale value classify the dataset.

In addition, in order to achieve the above object, the present invention also provides a data set classification device, the data set classification device includes: a memory, a processor, and a data set classification device stored on the memory and running on the processor A program that, when executed by the processor, implements the steps of the data set classification method described above.

In addition, in order to achieve the above object, the present invention also provides a computer storage medium on which a data set classification program is stored, and when the data set classification program is executed by a processor, the above-mentioned data set classification method is implemented A step of.

In the present invention, the data set to be classified is pre-divided into training set, test set and verification set, and a dynamic image is established based on the training set, the test set and the verification set; the adjustment received according to the dynamic image The instruction adjusts the training scale value corresponding to the training set to obtain the target training scale value; based on the target training scale value and the preset constraint conditions, the test scale value corresponding to the test set and the verification set are respectively adjusted. The corresponding validation scale value is updated, and the dataset is classified based on the target training scale value, the updated test scale value, and the updated validation scale value. By pre-dividing the data set into training set, test set and verification set, and establishing a dynamic image, the training scale value corresponding to the training set is adjusted in the dynamic image to obtain the target training scale value, and then according to the preset constraints. The test scale value corresponding to the test set and the verification scale value corresponding to the verification set are updated, and then the data set is divided, so that the user can reasonably divide the data set according to the dynamic image, which reduces the user's input cost and improves the performance. The user's understanding of the principles of data set classification also improves the efficiency and intelligence of data set classification, improves the efficiency of machine learning, and also improves the processing efficiency of data set classification.

Description of drawings

1 is a schematic diagram of a device structure of a hardware operating environment involved in an embodiment of the present invention;

2 is a schematic flowchart of a first embodiment of a data set classification method according to the present invention;

3 is a schematic diagram of a device module of a data set classification device of the present invention;

4 is a schematic diagram of an image scene in the data set classification method of the present invention;

FIG. 5 is a schematic diagram of the adjustment of the training set, the test set and the verification set in the data set classification method of the present invention.

The realization, functional characteristics and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments.

Detailed ways

It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

As shown in FIG. 1 , FIG. 1 is a schematic diagram of a device structure of a hardware operating environment involved in an embodiment of the present invention.

The data set classification device in the embodiment of the present invention may be a PC or a server device, on which a Java virtual machine runs.

As shown in FIG. 1 , the data set classification device may include: a processor 1001 , such as a CPU, a network interface 1004 , a user interface 1003 , a memory 1005 , and a communication bus 1002 . Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. Optionally, the network interface 1004 may include a standard wired interface and a wireless interface (eg, a WI-FI interface). The memory 1005 may be high-speed RAM memory, or may be non-volatile memory, such as disk memory. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001 .

Those skilled in the art can understand that the device structure shown in FIG. 1 does not constitute a limitation on the device, and may include more or less components than the one shown, or combine some components, or arrange different components.

As shown in FIG. 1 , the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a data set classification program.

In the device shown in FIG. 1 , the network interface 1004 is mainly used to connect to the background server and perform data communication with the background server; the user interface 1003 is mainly used to connect to the client (client) and perform data communication with the client; and the processor 1001 can be used to invoke the dataset classification program stored in memory 1005 and perform operations in the dataset classification method described below.

Based on the above hardware structure, an embodiment of the data set classification method of the present invention is proposed.

Referring to FIG. 2, FIG. 2 is a schematic flowchart of a first embodiment of a data set classification method according to the present invention. The method includes:

Step S10, pre-divide the data set to be classified into a training set, a test set and a verification set, and create a dynamic image based on the training set, the test set and the verification set;

Machine learning is the theory of how to make machines better at certain tasks. It can learn from data, rather than clearly coding the rules like traditional programming. In the process of machine learning, in order to make the trained machine learning algorithm model achieve good results in the actual environment, the data set for machine learning training is usually classified into training set, test set and validation set. Among them, the training set is a set of samples used for training, which is mainly used to train the parameters in the neural network. The validation set is a set of samples used to verify the performance of the model. After different neural networks are trained on the training set, the performance of each model is compared and judged through the validation set. Tests are used to objectively evaluate the performance of neural networks. A dataset is a collection of data.

In this embodiment, a data set to be classified (such as a data set to be subjected to machine learning) is obtained, and the data set is pre-divided into a training set, a test set and a verification set, that is, the data set is divided into three parts, and the training set is divided into three parts. Set, test set and validation set, these three parts are basically the same in essence, and they are all a kind of data set. In this embodiment, the data set can be randomly divided into training set, test set and verification set, and then the proportion of the training set in the data set, the proportion of the test set in the data set and the proportion of the verification set in the data set can be obtained. The proportions occupied, and dynamic images are built according to these proportions, so that users can visually see the classification of the dataset. The dynamic image can be a bar graph, a tree graph, etc., which is not limited here. For example, as shown in Figure 4, the proportion represented by the overall line is 100%, and the two sliders divide the entire line into three segments, and the length of each segment corresponds to the proportion of this dataset, that is, the training set occupies the proportion of the dataset. The proportion of the test set in the dataset is 80%, the proportion of the test set in the dataset is 10%, and the proportion of the validation set in the dataset is 10%. The proportion of each current data set is displayed under the line, such as 80 corresponding to the training set, 10 corresponding to the test set, and 10 corresponding to the validation set, so that the user can emotionally see the current data ratio during the adjustment process, so as to facilitate the user to analyze the data. division of the set. And different colors can be set for the line segments of different data sets in Figure 4 to reduce the cognitive cost of users.

Step S20, adjusting the training scale value corresponding to the training set according to the adjustment instruction received by the dynamic image, so as to obtain the target training scale value;

After establishing a dynamic image with the training scale value corresponding to the training set, the test scale value corresponding to the test set, and the verification scale value corresponding to the verification set, if it is detected that the user needs to adjust the training scale value, that is, the user receives the When the adjustment command is input, the pattern corresponding to the training scale value can be adjusted directly in the dynamic image, so that the user can intuitively see the adjustment effect, and after the adjustment of the training scale value is completed, the adjusted target training scale can be obtained. value to display the target training scale value in the dynamic image. Among them, the training proportion value is the proportion occupied by the training set in the data set. The test proportion value is the proportion of the test set in the dataset. The validation scale value is the fraction of the validation set in the dataset. The condition for triggering the adjustment instruction may be a number input by the user in the preset area of the dynamic image, or may be a touch operation performed on a preset button in the dynamic image, or the like.

Step S30, respectively update the test scale value corresponding to the test set and the verification scale value corresponding to the verification set based on the target training scale value and preset constraints, and update the target training scale value, update The subsequent test scale value and the updated validation scale value classify the dataset.

The preset constraints can be constraints set in advance, for example, the test scale value and the verification scale value are increased or decreased in equal proportions; or each time the test scale value increases by a preset number of scale values, the verification scale value increases by one percentage. or when the verification scale value increases by a preset number of scale values, the test scale value increases by one scale value, etc. It should be noted that, in this embodiment, no matter any changes are made to the training set, the test set and the validation set, the sum of the training set, the test set and the validation set must be equal to the data set. After obtaining the target training ratio value, it is also necessary to obtain the constraint condition set in advance, and then update the test ratio value corresponding to the test set and the verification ratio value corresponding to the verification set according to the constraint condition and the target training ratio value. After the update is completed, the data set is classified according to the target training ratio value, the updated test ratio value and the updated verification ratio value, and the data set is divided into the target training set, the target test set and the target verification set.

Moreover, it should be noted that when the user needs to adjust the test scale value and/or the verification scale value, the test scale value and/or the verification scale value can be adjusted directly, but the training scale value remains unchanged at this time. . For example, as shown in Figure 5, in the actual modeling process, the training set is necessary, the test set is more important, and the validation set can be used as an auxiliary when there is a test set, and its important arrangement is that the training set is larger than the test set , the test set is greater than or equal to the validation set, so different interactions can be set for the two sliders of the dynamic image in Figure 5. Slider 1 is used to adjust the proportion of the training set to the data set. When slider 1 is adjusted, the original division ratio between the test set and the validation set is maintained. The slider 2 is used to adjust the ratio between the test set and the validation set. Adjusting the slider 2 will not affect the proportion of the training set, that is, when the slider 2 is adjusted, the proportion of the training set occupying the data set remains unchanged, only Change the ratio of the test set to the dataset and the ratio of the validation set to the data set. For example, when the ratio between the training set, the test set and the validation set is 80:10:10, when the slider 1 is adjusted at this time, the slider 2 will also slide to keep the ratio between the test set and the validation set constant, that is The ratio between training set, test set and validation set was changed to 50:25:25. But when slider 2 is adjusted, the position of slider 1 remains unchanged, and only the ratio of test set and validation set is adjusted, that is, the ratio between training set, test set and validation set is adjusted from 80:10:10 to 80: 18:2.

And because the data set ratio under the line is linked with the slider, the two sliders will also adjust their positions according to the data after passing the acquaintance value. According to the importance of each data set, the three input boxes are divided into different numerical input limits. When inputting the ratio of the training set (assumed to be a), the values of the test set and the validation set are automatically adjusted to maintain the ratio between the two. The input range of a is 1≤a≤100, and the input of 100 means that the entire data is the training set. At this time, the test set The values of the set and validation set are both 0; when the ratio of the test set (assumed to be b) is input, it does not affect the ratio of the training set, and the value of the validation set is automatically adjusted. The input range of b is 1≤b≤100-a, and when 100-a is input Indicates that the entire data is divided into a training set and a validation set, and the value of the validation set is 0 at this time; when the ratio of the validation set is input (assumed to be c), it does not affect the ratio of the training set, and the value of the test set is automatically adjusted, and the input range of c is 1≤c <100-a.

By pre-dividing the data set to be performed machine learning into training set, test set and verification set, and establishing dynamic images based on the training set, test set and verification set; Adjusting the training scale value corresponding to the training set to obtain the target training scale value; based on the target training scale value and the preset constraint conditions, the test scale value corresponding to the test set and the verification scale corresponding to the verification set are respectively value is updated, and the dataset is classified based on the target training scale value, the updated test scale value, and the updated validation scale value. By pre-dividing the data set into training set, test set and verification set, and establishing a dynamic image, the training scale value corresponding to the training set is adjusted in the dynamic image to obtain the target training scale value, and then according to the preset constraints. The test scale value corresponding to the test set and the verification scale value corresponding to the verification set are updated, and then the data set is classified, so that the user can reasonably divide the data set according to the dynamic image, which reduces the user's input cost and improves the performance. The user's understanding of the principles of data set classification also improves the efficiency and intelligence of data set classification, improves the efficiency of machine learning, and also improves the processing efficiency of data set classification.

Further, based on the first embodiment of the data set classification method of the present invention, a second embodiment of the data set classification method of the present invention is proposed. This embodiment is step S10 of the first embodiment of the present invention, and the refinement of the step of creating a dynamic image based on the training set, test set and verification set includes:

Step a, obtaining the training ratio value of the training set in the data set, the test ratio value of the test set in the data set and the verification ratio value of the verification set in the data set;

After pre-dividing the data set into training set and test set validation set, it is also necessary to calculate the proportion occupied by the training set in the data set, namely the training proportion value; the proportion occupied by the test set in the data set, namely the test proportion value; The proportion that the set occupies in the data set, that is, the validation proportion value.

In step b, the training scale value, the test scale value and the verification scale value are transferred into a preset image to obtain a dynamic image.

Obtain the preset image (that is, the original image without any parameters, and the original image needs to reflect the difference of at least three different elements), and then input the previously calculated training scale value, test scale value and verification scale value into the preset image. Set the image in the image, so as to obtain a dynamic image, so that the user can more intuitively see that the data set is divided into training set, test set and verification set.

In this embodiment, a dynamic image is obtained by passing the training scale value, the test scale value and the verification scale value into the preset image, thereby ensuring that the user can intuitively see that the data set is divided into training set, test set and The rationale for the validation set.

Further, the step of respectively updating the test scale value corresponding to the test set and the verification scale value corresponding to the verification set based on the target training scale value and the preset constraint conditions includes:

Step c, calculating the remaining ratio value of the data set according to the target training ratio value, and respectively corresponding to the test ratio value of the test set and the verification set based on preset constraints and the remaining ratio value. The validation scale value is updated.

After the target training ratio value is obtained, the remaining ratio value of the data set is obtained by subtracting the target training ratio value from the data value carried by the data set itself. For example, when the data value carried by the dataset itself is 100, when the adjusted target training ratio value is 80, then 20 can be obtained by subtracting 80 from 100, where 20 is the remaining ratio value. Then, the preset constraint conditions are obtained, and the test scale value corresponding to the test set and the verification scale value corresponding to the verification set are updated according to the remaining scale value. And, the sum of the updated test scale value and the updated verification scale value should be equal to the remaining scale value. For example, when the test scale value is 20, the verification scale value is 20, and the remaining scale value is also 20, and the constraint condition is proportionally increasing or decreasing, then the updated test scale value is 10, and the updated verification scale value is Just 10.

In this embodiment, the test scale value and the verification scale value are updated according to the remaining scale value and constraint conditions of the data set, thereby ensuring the accuracy of the updated test scale value and the verification scale value.

Specifically, the steps of respectively updating the test scale value corresponding to the test set and the verification scale value corresponding to the verification set based on the preset constraint condition and the remaining scale value include:

Step c1, obtaining the first sum value between the test scale value corresponding to the test set and the verification scale value corresponding to the verification set, and judging whether the first sum value is greater than the remaining scale value;

In this embodiment, the constraints include proportional adjustment.

Obtain the test proportion value corresponding to the test set and the verification proportion value corresponding to the verification set, and calculate the sum value between the test proportion value and the verification proportion value, use this sum value as the first sum value, and then judge whether the first sum value is greater than the remaining scale value. Then perform different operations based on different judgment results.

Step c2, if it is less than, perform equal-proportional increase and update processing on the test proportional value and the verification proportional value based on a preset equal-proportional adjustment, until the first sum value is equal to the remaining proportional value.

When it is judged that the first sum value is smaller than the remaining proportion value, and the constraint condition is equal proportion adjustment, the test proportion value and the verification proportion value are incrementally processed in equal proportions at the same time, until the incremental test proportion value and the verification proportion value are processed. When the sum value is equal to the remaining scale value, the update processing of the test scale value and the verification scale value is stopped.

In this embodiment, when the first sum of the test scale value and the verification scale value is greater than the remaining scale value, the test scale value and the verification scale value are increased in equal proportions at the same time, thereby ensuring the updated test scale value and verify the accuracy of the scale values.

Specifically, after the step of judging whether the first sum value is greater than the remaining ratio value, the step includes:

Step c3, if it is greater than the value, perform equal proportional reduction and update processing on the test proportional value and the verification proportional value at the same time based on a preset equal proportional adjustment, until the first sum value is equal to the remaining proportional value.

When it is determined that the first sum value is greater than the remaining proportional value, and the constraint condition is equal-proportional adjustment, the test proportional value and the verification proportional value are simultaneously decreased in equal proportions until the test proportional value and the verification proportional value after the decrease processing are processed. When the sum value is equal to the remaining scale value, the update processing of the test scale value and the verification scale value is stopped.

In this embodiment, when the first sum of the test scale value and the verification scale value is smaller than the remaining scale value, the test scale value and the verification scale value are simultaneously reduced in equal proportions, thereby ensuring the updated test scale value and verify the accuracy of the scale value.

Further, based on any one of the first to second embodiments of the data set classification method of the present invention, a third embodiment of the data set classification method of the present invention is proposed. This embodiment is step S10 of the first embodiment of the present invention. After the step of establishing a dynamic image based on the training set, the test set and the verification set, it includes:

Step e, obtaining the training scale value corresponding to the training set in the dynamic image, the test scale value corresponding to the test set and the verification scale value corresponding to the verification set;

After the dynamic image is established, it is necessary to obtain the training scale value corresponding to the training set, the test scale value corresponding to the test set, and the verification scale value corresponding to the verification set in the dynamic image.

Step f, receiving a test scale value update instruction, and updating the test scale value based on the test scale value update instruction to obtain an update test scale value;

A test scale value update instruction for updating the test scale value input by the user is received, and the test scale value is updated according to the test scale value update instruction to obtain the updated update test scale value. And when the test scale value is updated, the data of its training scale value remains unchanged.

Step g, the verification ratio value is updated according to the training ratio value and the update test ratio value to obtain a new verification ratio value, and based on the training ratio value, the update test ratio value and the new verification ratio value. Validation scale values classify the dataset.

After the update test scale value is obtained, since the data of the training scale value remains unchanged, it is also necessary to subtract the update test scale value and the training scale value from the data value of the data set to obtain the remaining value of the data set, and according to The residual value adjusts the verification scale value until the verification scale value is equal to the residual value, then the verification scale value at this time is used as the new verification scale value. Then, the dataset is classified according to the training scale value, the update test scale value and the new verification scale value to obtain the actual training value, test value and scale value.

In this embodiment, the test scale value is updated according to the test scale value update instruction to obtain the updated test scale value, and then the verification scale value is updated by updating the test scale value and the training scale value, and then the data set is classified, Therefore, the user can reasonably divide the data set according to the dynamic image, which reduces the input cost of the user and improves the user's understanding of the classification principle of the data set.

Further, after the step of establishing a dynamic image based on the training set, the test set and the verification set, it includes:

Step v, obtaining the training scale value corresponding to the training set in the dynamic image, the test scale value corresponding to the test set and the verification scale value corresponding to the verification set;

After the dynamic image is established, it is necessary to obtain the training scale value corresponding to the training set, the test scale value corresponding to the test set, and the verification scale value corresponding to the verification set in the dynamic image.

Step h, receiving a verification ratio value update instruction, and updating the verification ratio value based on the verification ratio value update command to obtain a new verification ratio value;

Receive a verification scale value update instruction for updating the test scale value input by the user, and perform update processing on the verification scale value according to the verification scale value update instruction to obtain an updated new verification scale value. And when the validation scale value is updated, the data of its training scale value remains unchanged.

Step k, update the test scale value based on the training scale value and the new verification scale value to obtain a new test scale value, and based on the training scale value, the new verification scale value and all The dataset is classified using the new test scale value.

After the new verification scale value is obtained, since the data of the training scale value remains unchanged, it is also necessary to subtract the new verification scale value and the training scale value from the data value of the data set to obtain the remaining value of the data set, And adjust the test scale value according to the residual value until the test scale value is equal to the residual value, then take the test scale value at this time as the new test scale value. The dataset is then classified according to the training scale value, the new test scale value and the new validation scale value to obtain the actual training value, test value and scale value.

In this embodiment, the verification ratio value is updated according to the verification ratio value update instruction to obtain a new verification ratio value, and then the test ratio value is updated through the new verification ratio value and the training ratio value, and then the data set is updated. Classification, so that the user can reasonably divide the data set according to the dynamic image, reduce the input cost of the user, and improve the user's understanding of the classification principle of the data set.

The present invention also provides a data set classification device. Referring to FIG. 3, the data set classification device includes:

establishing a module for pre-dividing the data set to be classified into a training set, a test set and a verification set, and establishing a dynamic image based on the training set, the test set and the verification set;

an acquisition module, configured to adjust the training scale value corresponding to the training set according to the adjustment instruction received by the dynamic image, so as to obtain the target training scale value;

A classification module, configured to update the test scale value corresponding to the test set and the verification scale value corresponding to the verification set based on the target training scale value, and based on the target training scale value, the updated test scale value The scale value and the updated validation scale value classify the dataset.

Optionally, the establishment module is also used for:

Obtain the training ratio value of the training set in the data set, the test ratio value of the test set in the data set and the verification ratio value of the verification set in the data set;

The training scale value, the test scale value and the verification scale value are transferred into a preset image to obtain a dynamic image.

Optionally, the classification module is also used for:

Calculate the remaining ratio value of the data set according to the target training ratio value, and based on preset constraints and the remaining ratio value, respectively, the test ratio value corresponding to the test set and the verification ratio corresponding to the verification set value is updated.

Optionally, the constraints include proportional adjustment, and the classification module is further configured to:

Obtain the first sum value between the test scale value corresponding to the test set and the verification scale value corresponding to the verification set, and determine whether the first sum value is greater than the remaining scale value;

If it is less than, the test proportional value and the verification proportional value are simultaneously increased and updated in equal proportions based on a preset equal proportional adjustment, until the first sum value is equal to the remaining proportional value.

Optionally, the classification module is also used for:

If it is greater than the value, the test proportional value and the verification proportional value are simultaneously subjected to proportional reduction and update processing based on a preset equal proportional adjustment, until the first sum value is equal to the remaining proportional value.

Optionally, the data set classification device further includes:

Obtain the training scale value corresponding to the training set in the dynamic image, the test scale value corresponding to the test set and the verification scale value corresponding to the verification set;

receiving a test scale value update instruction, and updating the test scale value based on the test scale value update instruction to obtain an updated test scale value;

The verification scale value is updated according to the training scale value and the update test scale value to obtain a new verification scale value, and based on the training scale value, the update test scale value and the new verification scale value Classify the dataset.

Optionally, the data set classification device further includes:

Obtain the training scale value corresponding to the training set in the dynamic image, the test scale value corresponding to the test set and the verification scale value corresponding to the verification set;

receiving a verification scale value update instruction, and updating the verification scale value based on the verification scale value update instruction to obtain a new verification scale value;

The test scale value is updated based on the training scale value and the new validation scale value to obtain a new test scale value, and based on the training scale value, the new validation scale value and the new test The scale value classifies the dataset.

For the methods executed by the above program modules, reference may be made to the various embodiments of the data set classification method of the present invention, which will not be repeated here.

The present invention also provides a computer storage medium.

A data set classification program is stored on the computer storage medium of the present invention, and when the data set classification program is executed by the processor, the steps of the above-mentioned data set classification method are implemented.

For the method implemented when the data set classification program running on the processor is executed, reference may be made to the various embodiments of the data set classification method of the present invention, which will not be repeated here.

It should be noted that, herein, the terms "comprising", "comprising" or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article or system comprising a series of elements includes not only those elements, It also includes other elements not expressly listed or inherent to such a process, method, article or system. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in the process, method, article or system that includes the element.

The above-mentioned serial numbers of the embodiments of the present invention are only for description, and do not represent the advantages or disadvantages of the embodiments.

From the description of the above embodiments, those skilled in the art can clearly understand that the method of the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is better implementation. Based on such understanding, the technical solutions of the present invention can be embodied in the form of software products in essence or the parts that make contributions to the prior art, and the computer software products are stored in a storage medium (such as ROM/RAM) as described above. , magnetic disk, optical disk), including several instructions to make a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the methods described in the various embodiments of the present invention.

The above are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the description and drawings of the present invention, or directly or indirectly applied in other related technical fields , are similarly included in the scope of patent protection of the present invention.

Claims (10)

1. a data set classification method, is characterized in that, described data set classification method comprises the steps: Pre-dividing the data set to be classified into a training set, a test set and a verification set, and creating a dynamic image based on the training set, the test set and the verification set; Adjust the training scale value corresponding to the training set according to the adjustment instruction received from the dynamic image to obtain the target training scale value; The test scale value corresponding to the test set and the verification scale value corresponding to the verification set are respectively updated based on the target training scale value and preset constraints, and based on the target training scale value, the updated all scale values The data set is classified according to the test scale value and the updated validation scale value. 2. The data set classification method as claimed in claim 1, wherein the step of establishing a dynamic image based on the training set, the test set and the verification set comprises: Obtain the training ratio value of the training set in the data set, the test ratio value of the test set in the data set and the verification ratio value of the verification set in the data set; The training scale value, the test scale value and the verification scale value are transferred into a preset image to obtain a dynamic image. 3. The data set classification method according to claim 1, wherein the test proportion value corresponding to the test set and the verification set corresponding to the test proportion value based on the target training proportion value and the preset constraint condition are respectively The steps for updating the validation scale value include: Calculate the remaining ratio value of the data set according to the target training ratio value, and based on preset constraints and the remaining ratio value, respectively, the test ratio value corresponding to the test set and the verification ratio corresponding to the verification set value is updated. 4. data set classification method as claimed in claim 3, is characterized in that, described constraint condition comprises equal proportion adjustment, The step of respectively updating the test scale value corresponding to the test set and the verification scale value corresponding to the verification set based on the preset constraint condition and the remaining scale value includes: Obtain the first sum value between the test scale value corresponding to the test set and the verification scale value corresponding to the verification set, and determine whether the first sum value is greater than the remaining scale value; If it is less than, the test proportional value and the verification proportional value are simultaneously increased and updated in equal proportions based on a preset equal proportional adjustment, until the first sum value is equal to the remaining proportional value. 5. The data set classification method according to claim 4, wherein after the step of judging whether the first sum value is greater than the remaining ratio value, the method comprises: If it is greater than the value, the test proportional value and the verification proportional value are simultaneously subjected to proportional reduction and update processing based on a preset equal proportional adjustment, until the first sum value is equal to the remaining proportional value. 6. The data set classification method according to any one of claims 1-5, wherein after the step of establishing a dynamic image based on the training set, the test set and the verification set, the method further comprises: Obtain the training scale value corresponding to the training set in the dynamic image, the test scale value corresponding to the test set and the verification scale value corresponding to the verification set; receiving a test scale value update instruction, and updating the test scale value based on the test scale value update instruction to obtain an updated test scale value; The verification scale value is updated according to the training scale value and the update test scale value to obtain a new verification scale value, and based on the training scale value, the update test scale value and the new verification scale value Classify the dataset. 7. The data set classification method according to any one of claims 1-5, wherein after the step of establishing a dynamic image based on the training set, the test set and the verification set, the method further comprises: Obtain the training scale value corresponding to the training set in the dynamic image, the test scale value corresponding to the test set and the verification scale value corresponding to the verification set; receiving a verification scale value update instruction, and updating the verification scale value based on the verification scale value update instruction to obtain a new verification scale value; The test scale value is updated based on the training scale value and the new validation scale value to obtain a new test scale value, and based on the training scale value, the new validation scale value and the new test The scale value classifies the dataset. 8. A data set classification device, wherein the data set classification device comprises: establishing a module for pre-dividing the data set to be classified into a training set, a test set and a verification set, and establishing a dynamic image based on the training set, the test set and the verification set; an acquisition module, configured to adjust the training scale value corresponding to the training set according to the adjustment instruction received by the dynamic image, so as to obtain the target training scale value; A classification module, configured to update the test scale value corresponding to the test set and the verification scale value corresponding to the verification set based on the target training scale value, and based on the target training scale value, the updated test scale value The scale value and the updated validation scale value classify the dataset. 9. A data set classification device, characterized in that the data set classification device comprises: a memory, a processor, and a data set classification program stored in the memory and executable on the processor, the data A set classification program, when executed by the processor, implements the steps of a data set classification method as claimed in any one of claims 1 to 7. 10. A computer storage medium, wherein a data set classification program is stored on the computer storage medium, and when the data set classification program is executed by a processor, the method according to any one of claims 1 to 7 is implemented. Steps of a dataset classification method.

Images