CN111738762A - Method, device, equipment and storage medium for determining recovery price of poor assets - Google Patents
Method, device, equipment and storage medium for determining recovery price of poor assets Download PDFInfo
- Publication number
- CN111738762A CN111738762A CN202010565628.1A CN202010565628A CN111738762A CN 111738762 A CN111738762 A CN 111738762A CN 202010565628 A CN202010565628 A CN 202010565628A CN 111738762 A CN111738762 A CN 111738762A
- Authority
- CN
- China
- Prior art keywords
- target
- asset
- bad
- assets
- recovery
- 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
Links
Images
Classifications
-
- 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
- G06Q30/0283—Price estimation or determination
-
- 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
- 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
- G06Q40/03—Credit; Loans; Processing thereof
-
- 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
- G06Q40/06—Asset management; Financial planning or analysis
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Technology Law (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Human Resources & Organizations (AREA)
- Operations Research (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application relates to a method, a device, equipment and a storage medium for determining a recovery price of a poor asset. The method comprises the following steps: acquiring characteristic data of target bad assets; inputting the characteristic data into a preset classification model to obtain a classification result of the target bad asset; when the classification result is not equal to a target classification result, inputting the characteristic data into a preset prediction model to obtain the recovery rate of the target bad asset, wherein the target classification result is related to the attribution attribute of the target bad asset; and determining the recovery price of the target bad assets according to the recovery rate. The method can improve the pricing efficiency of the poor assets and the accuracy of pricing results.
Description
Technical Field
The application relates to the field of internet, in particular to a method, a device, equipment and a storage medium for determining a recovery price of a poor asset.
Background
The poor property refers to the property of the bank which is in a non-good operation state, cannot bring normal interest income to the bank in time and even cannot recover principal in time, such as poor loan and the like. When poor assets appear in bank assets, how to price the poor assets, namely, determining the recovery price of the poor assets is a technical problem to be solved urgently by technical personnel in the field.
Traditional bad asset pricing is mainly based on off-line pricing. In particular, it is common for an asset retention manager to artificially empirically price undesirable assets based on field investigation results for the undesirable assets. However, the traditional method has poor timeliness, so that the pricing efficiency is low, and the pricing result is excessively dependent on the experience and personal quality of an asset security manager, so that the pricing result is inaccurate.
Disclosure of Invention
In view of the above, it is necessary to provide a method, an apparatus, a device and a storage medium for determining a recycling price of a poor asset, aiming at the technical problems of the conventional technology that the pricing efficiency of the poor asset is low and the pricing result is inaccurate.
In a first aspect, an embodiment of the present application provides a method for determining a recovery price of an undesirable asset, including:
acquiring characteristic data of target bad assets;
inputting the characteristic data into a preset classification model to obtain a classification result of the target bad asset;
when the classification result is not equal to a target classification result, inputting the characteristic data into a preset prediction model to obtain the recovery rate of the target bad asset, wherein the target classification result is related to the attribution attribute of the target bad asset;
and determining the recovery price of the target bad assets according to the recovery rate.
In a second aspect, an embodiment of the present application provides an apparatus for determining a recovery price of an undesirable asset, including:
the acquisition module is used for acquiring the characteristic data of the target bad asset;
the classification module is used for inputting the characteristic data into a preset classification model to obtain a classification result of the target bad asset;
the recovery rate prediction module is used for inputting the characteristic data into a preset prediction model to obtain the recovery rate of the target bad asset when the classification result is not equal to a target classification result, wherein the target classification result is related to the attribution attribute of the target bad asset;
and the first determining module is used for determining the recovery price of the target bad asset according to the recovery rate.
In a third aspect, an embodiment of the present application provides a computer device, including a memory and a processor, where the memory stores a computer program, and the processor implements the method for determining a recovery price of a poor asset provided by the first aspect of the embodiment of the present application when executing the computer program.
In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method for determining a recovery price of an undesirable asset provided by the first aspect of the embodiment of the present application.
According to the method, the device, the equipment and the storage medium for determining the recovery price of the poor assets, after the characteristic data of the target poor assets are obtained, the computer equipment inputs the characteristic data into the preset classification model to obtain the classification result of the target poor assets, when the classification result is not equal to the target classification result, the characteristic data are input into the preset prediction model to obtain the recovery rate of the target poor assets, and the recovery price of the target poor assets is determined according to the obtained recovery rate. In the whole pricing process of the target bad assets, the computer equipment can predict the recovery rate of the target bad assets based on the pre-trained classification model and the prediction model and determine the recovery price of the target bad assets based on the predicted recovery rate, namely, the whole pricing process is automatically completed by the computer equipment, and the asset security manager is not required to perform offline pricing based on own experience, so that the pricing efficiency of the bad assets is improved, and meanwhile, compared with manual pricing, the technical scheme also improves the accuracy of pricing results.
Drawings
FIG. 1 is a schematic flow chart of a method for determining a recovery price of an undesirable asset according to an embodiment of the present application;
fig. 2 is a schematic flowchart of an obtaining process of a classification model according to an embodiment of the present application;
FIG. 3 is a flowchart illustrating a process of obtaining a prediction model according to an embodiment of the present disclosure;
fig. 4 is a schematic structural diagram of a device for determining a recovery price of an undesirable asset according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a computer device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions in the embodiments of the present application are further described in detail by the following embodiments in combination with the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
It should be noted that the execution subject of the method embodiments described below may be a device for determining the recovery price of the undesirable asset, and the device may be implemented as part or all of a computer device by software, hardware or a combination of software and hardware. Optionally, the computer device may be an electronic device that has a data processing function and can interact with an external device or a user, such as a personal computer pc (personal computer), a mobile terminal, and a portable device, and of course, the computer device may also be a server, and the specific form of the computer device is not limited in this embodiment. The method embodiments described below are described by way of example with the execution subject being a computer device.
Fig. 1 is a schematic flow chart of a method for determining a recovery price of an undesirable asset according to an embodiment of the present application. This embodiment relates to a specific process of how a computer device can price a targeted undesirable asset. As shown in fig. 1, the method may include:
and S101, acquiring characteristic data of the target bad assets.
Specifically, the bad assets refer to bank assets which are in a non-good operation state, cannot bring normal interest income to the bank in time and even cannot withdraw principal money in time. The bank assets mainly include funds, securities investments, client loans, securities purchased under the agreement of peer release and repurchase, and other assets (fixed assets such as property and facilities), and when the bank assets are in bad condition, the part of the assets are defined as bad assets. A target assets is assets that are not good for pricing.
The characteristic data of the target undesirable asset may include: basic contract information, basic enterprise information, running contract information, financial enterprise information, collateral information, basic personal information, mortisable asset information and the like. The basic contract information refers to various basic information during contract signing, such as contract signing time, asset amount related to the contract and the like; the basic information of the enterprise mainly comprises basic conditions of the enterprise, such as registration time, registration amount and the like; the contract flow information mainly comprises the change condition of the repayment amount or loan balance data of the contract in a certain period; the enterprise financial information mainly comprises financial condition information of the enterprise when signing a contract, such as turnover, profit margin and the like; the information of the extruded material mainly includes the basic condition of the extruded material, such as the type of extruded material and the estimated value of extruded material; the personal basic information mainly comprises basic conditions of the person, such as income, age, sex, school calendar and the like; the debtable asset information mainly includes information of the condition of the debtor's debtable asset, such as the category of the debt's assets and the valuation of the debt's assets. Of course, in practical applications, different target bad assets may correspond to different feature data, such as individual loan non-mortgage type bad assets, and the feature data does not include mortgage information.
The characteristic data of the target undesirable asset may be stored in a storage device of the computer device in advance, and when pricing of the target undesirable asset is required, the computer device may select the characteristic data of the target undesirable asset from the storage device. Of course, the computer device may also obtain the characteristic data of the target bad asset from other external devices. For example, the feature data of the target undesirable asset is stored in the cloud, and when the target undesirable asset needs to be priced, the computer device can acquire the feature data of the target undesirable asset from the cloud.
And S102, inputting the characteristic data into a preset classification model to obtain a classification result of the target bad asset.
Specifically, the classification model may include any one of LightGBM (gradient Boosting framework, using a learning algorithm based on a decision tree), XGBoost (extreme gradient Boosting), and random forest. The classification model may classify the target defective asset into zero recovery or non-zero recovery, but may classify the target defective asset into full recovery or non-full recovery, or may classify the target defective asset into another classification result. Namely, the corresponding classification model can be selected to classify the target bad assets according to the actual application requirements. Optionally, when the attribution attribute of the target bad asset is an enterprise-class bad asset, particularly a small enterprise-class bad asset, since more than 70% of bad debt items in the small enterprise-class bad asset are zero-reclaimed, the feature data of the target bad asset may be input to the corresponding classification model to classify the target bad asset as zero-reclaimed or non-zero reclaimed. When the attribution attribute of the target bad asset is personal bad assets, particularly personal non-mortgage loan bad assets, more than 95% of bad debt items in the personal non-mortgage loan bad assets can be completely recycled, so that the characteristic data of the target bad asset can be input into the corresponding classification model to classify the target bad asset into complete recycling or incomplete recycling.
Illustratively, taking the classification model as a random forest as an example, the random forest model is formed by combining a series of mutually independent decision trees, and each decision tree forms the minimum component of the whole random forest model. Its expression form can be written as
Where RF (x) is the output of the random forest model, fiThe method comprises the steps of representing a single decision tree classifier model in a random forest model, representing the output of the single decision tree classifier model by Y, representing an indicative function by I, and representing the number of decision trees in the random forest model by b. And after an independent variable x of the random forest model is given, each decision tree independently judges the input without mutual influence, and finally the classification result of the whole classifier is selected through voting. The decision-making capability of a single decision-making tree is weak, but the decision-making capability of a series of decision-making trees is strong when the decision-making trees are organically integrated.
And continuously taking the classification model as a random forest, taking the classification model as an example for classifying the target bad assets into complete recovery or incomplete recovery, inputting the feature data of the target bad assets into the trained random forest model by the computer equipment, classifying the feature data of the target bad assets by each decision tree in the random forest model to obtain the classification result of each decision tree, then simply voting the classification results, and selecting the final classification result according to the voting. Wherein the final classification result is that the target bad assets are completely recycled or not completely recycled.
Certainly, the LightGBM has the advantages of faster training efficiency, low memory occupancy rate, higher classification accuracy, support for parallelization learning, capability of processing large-scale data and the like, so the LightGBM can be used as the classification model to classify the target bad assets.
S103, when the classification result is not equal to the target classification result, inputting the characteristic data into a preset prediction model to obtain the recovery rate of the target bad asset.
Specifically, the target classification result is related to the attribution attribute of the target bad asset. The attribution attributes of the target assets may include enterprise-type assets or personal-type assets. Optionally, when the attribution attribute of the target bad asset is an enterprise type bad asset, the target classification result is zero recovery; and when the attribution attribute of the target bad asset is personal bad assets, the target classification result is complete recovery. The above-mentioned zero recovery and full recovery belong to extreme recovery, and the non-zero recovery and non-full recovery belong to non-extreme recovery, that is, when the classification result of the target bad asset does not belong to extreme recovery, the computer device may input the characteristic data of the target bad asset into the prediction model to further determine the recovery rate of the target bad asset in case of non-extreme recovery.
Taking the attribution attribute of the target bad asset as the personal bad asset as an example, the computer equipment inputs the characteristic data of the personal bad asset into the corresponding classification model so as to predict whether the personal bad asset can be completely recycled. When the classification result of the personal type poor asset is determined to be not completely recovered, the computer equipment inputs the characteristic data of the personal type poor asset into a prediction model so as to further predict the recovery rate of the personal type poor asset. When the classification result of the personal type poor assets is determined to be completely recovered, the next prediction through a prediction model is not needed, and the recovery rate of the personal type poor assets can be directly determined to be 100%.
It will be appreciated that the computer device classifies the target undesirable asset via a classification model, and that the classification results may direct the asset retention manager to manage the target undesirable asset. When the target bad asset is determined to be the bad asset with zero recovery, the target bad asset can not be managed any more, the management cost of the target bad asset is reduced, and the disposal rate of the bad asset is improved. In addition, when the classification result of the target bad asset is not equal to the target classification result, the computer device further predicts the recovery rate of the target bad asset through the prediction model, and then can further process the target bad asset based on the recovery rate as a reference basis for asset processing, such as collection urging, verification and sale of assets or package sale of assets and the like according to the recovery rate.
And S104, determining the recovery price of the target bad asset according to the recovery rate.
Specifically, after obtaining the recovery rate for the target asset, the computer device may further determine a recovery price for the target asset based on the recovery rate. Optionally, the process of S104 may be: acquiring the credit balance of the target bad assets; and determining the recovery price of the target bad assets according to the recovery rate and the credit balance. Wherein the claim balance is the difference between the total contract amount and the collected amount of the target bad asset. Taking target bad property as an example of bad loan, the computer device obtains the total loan amount (including principal and interest) and the recovered amount (including recovered principal and interest) of the bad loan, determines the creditor balance of the bad loan according to the difference between the total loan amount and the recovered amount, and determines the recovery price of the bad loan based on the product of the creditor balance and the determined recovery rate.
Optionally, when the classification result of the target bad asset is equal to the target classification result (i.e., zero recovery or full recovery), the computer device may determine the recovery price of the target bad asset based on the recovery rate corresponding to the target classification result.
For example, when the attribution attribute of the target bad asset is the enterprise-class bad asset, if the computer device classifies the enterprise-class bad asset through the classification model, and the obtained classification result is zero recovery, and the recovery rate of the enterprise-class bad asset is 0%, the recovery price of the enterprise-class bad asset is determined to be 0 based on the recovery rate. When the attribution attribute of the target bad asset is personal bad assets, if the personal bad assets are classified by the computer equipment through the classification model, the obtained classification result is complete recovery, and the recovery rate of the personal bad assets is 100%, determining that the recovery price of the personal bad assets is equal to the balance of the creditor thereof based on the recovery rate.
Optionally, in order to facilitate the processing of the feature data by the classification model and the prediction model, on the basis of the foregoing embodiment, optionally before the foregoing S102, the method may further include: and preprocessing the characteristic data according to the data type of the characteristic data.
The data types of the feature data may include a value class, a date class, a type class, and a description class. For the above various types of feature data, the following process may be performed:
1) numerical class
The feature data of the numerical class may be automatically identified by the corresponding storage format by looking up the fields in numpy.flow64 or numpy.int64 format. It should be noted that, sometimes fields with non-numeric meanings, such as numbers and codes, are stored in the above two numeric formats, so it is necessary to check each numeric field to determine whether it is a real meaningful value.
2) Dates and the like
The characteristic data of the date class can be automatically identified by the corresponding storage format by looking up the fields of the pandas. It should be noted that the date field may have a string format date in a non-standard storage format, such as a string "xxxx-xx-xx" format, a date and time mixed format, etc. At this time, the dates of the non-standard storage formats need to be identified and extracted according to the regular expressions.
3) Class of type
The characteristic data of the type class can be preprocessed by using single hot coding or multi-hot coding, namely, the fields containing multiple types are coded one by one according to a corresponding coding mode.
4) Description class
And the field is generally more than 30 characters and can be judged to belong to the description class field. The processing of the description field is complex, and if the model precision meets the requirement, the description characteristic data can be deleted according to the requirement; if the model accuracy is further improved, the long character string may be mapped to a numerical value (e.g., to a length of the character string or a frequency of occurrence of keywords, etc.).
According to the method for determining the recovery price of the poor assets, after the feature data of the target poor assets are obtained, the computer equipment inputs the feature data into a preset classification model to obtain the classification result of the target poor assets, when the classification result is not equal to the target classification result, the feature data are input into a preset prediction model to obtain the recovery rate of the target poor assets, and the recovery price of the target poor assets is determined according to the obtained recovery rate. In the whole pricing process of the target bad assets, the computer equipment can predict the recovery rate of the target bad assets based on the pre-trained classification model and the prediction model and determine the recovery price of the target bad assets based on the predicted recovery rate, namely, the whole pricing process is automatically completed by the computer equipment, and the asset security manager is not required to perform offline pricing based on own experience, so that the pricing efficiency of the bad assets is improved, and meanwhile, compared with manual pricing, the technical scheme also improves the accuracy of pricing results.
In one embodiment, an obtaining process of the classification model is further provided. Optionally, the classification model is obtained by performing model training based on a historical undesirable asset classification set and a preset first basic model, where the historical undesirable asset classification set includes feature data of historical undesirable assets and an actual classification result.
The specific model training process may refer to the process shown in fig. 2, and specifically, before S101, the method further includes:
s201, performing model training according to a historical bad asset classification set and a preset first basic model, and determining an actual value of a classification parameter of the first basic model when a loss value of a loss function reaches a preset threshold value and is kept stable; wherein the first basic model comprises an initial value of a classification parameter.
Specifically, the computer device may perform data preprocessing and data derivation on the historical undesirable asset classification sets prior to model training. In order to make the training process more comprehensive and make the robustness of the classification model obtained by training higher, the computer device can select various types of feature data, and at this time, the feature data needs to be preprocessed according to the types of the feature data. The preprocessing process of the specific training data may refer to the preprocessing process of the feature data of the target bad asset, and this embodiment is not described herein again.
To increase the amount of data in the historical undesirable asset classification set employed in the model training process, the computer device may perform data derivation on the historical undesirable asset classification set. Taking date derivation as an example, since the data entered into the first base model has a large number of date fields, feature derivation of the emphasis of the date fields may be considered. The date derivation method mainly comprises the following steps: the date is subject to subtraction. Meanwhile, the date derivation method is divided into the following two methods according to whether the date has business significance or not:
1) business mean date difference
And performing difference on a specific date field to serve as a derivative variable, wherein the difference is required to contain business significance and can be interpreted by business.
2) Difference of full arranged date
And in the case that the number of date data is not too large, performing difference on every two full date fields to serve as a derivative variable. It should be noted that, by using the data with the full-range date difference, the number of the date data cannot be too large, otherwise, the dimension of the date difference almost occupies the wide table, so that the weight of the date class is too high, and the weight of other data is too low, which causes the problem of overfitting of the trained model.
Optionally, the first basic model may be LightGBM, XGBoost, random forest, or other types of models. As training progresses, the loss value of the loss function maintains a stable value, and therefore model convergence is achieved.
S202, replacing the initial value of the classification parameter in the first basic model with the actual value of the classification parameter to obtain the classification model.
Optionally, after obtaining the classification model, the classification model may be further subjected to model evaluation. In practical application, AUC (Area Under Curve, defined as the Area enclosed by the coordinate axis Under the ROC Curve), accuracy, precision, confusion matrix, and the like of the classification model can be calculated respectively, the classification effect of the classification model is evaluated based on the calculation result of each evaluation index, and the classification parameters of the classification model are adjusted locally according to the evaluation result.
In one embodiment, a process for obtaining the above-mentioned predictive model is also provided. On the basis of the foregoing embodiment, optionally, the prediction model is obtained by performing model training based on a historical undesirable asset prediction set and a preset second base model, where the historical undesirable asset prediction set includes characteristic data and an actual recovery rate of the historical undesirable asset.
The specific model training process may refer to the process shown in fig. 3, and specifically, before S101, the method further includes:
s301, performing model training according to the historical bad asset prediction set and a preset second basic model, and determining the actual value of the prediction parameter of the second basic model when the loss value of the loss function reaches a preset threshold value and is kept stable; wherein the second base model comprises initial values of the prediction parameters.
In particular, the computer device may perform data preprocessing and data derivation on the historical bad asset prediction set prior to model training. Regarding the process of the computer device for performing data preprocessing and data derivation on the historical bad asset prediction set, reference may be made to the process of the computer device for performing data preprocessing and data derivation on the historical bad asset classification set, and this embodiment is not described herein again.
Optionally, the second basic model may be LightGBM, XGBoost, random forest, or other types of models. As training progresses, the loss value of the loss function maintains a stable value, and therefore model convergence is achieved.
S302, replacing the initial value of the prediction parameter in the second basic model with the actual value of the prediction parameter to obtain the prediction model.
Optionally, after obtaining the prediction model, model evaluation may be performed on the prediction model. In practical application, Root Mean Square Error (RMSE), average Error rate, maximum Error rate, R-Square value, etc. of the prediction model may be calculated respectively, the prediction effect of the prediction model may be evaluated based on the calculation result of each evaluation index, and the prediction parameters of the prediction model may be locally adjusted according to the evaluation result.
In this embodiment, the classification model used in the pricing process of the target undesirable asset is obtained by performing model training based on a large number of historical undesirable asset classification sets and the first basic model, and the used prediction model is obtained by performing model training based on a large number of historical undesirable asset prediction sets and the second basic model, and the first basic model and the second basic model are machine learning models, so that the robustness of the classification model and the prediction model obtained by training is higher, and the accuracy of the pricing result of the target undesirable asset is further improved.
Fig. 4 is a schematic structural diagram of a device for determining a recovery price of a poor asset according to an embodiment of the present application. As shown in fig. 4, the apparatus may include: an acquisition module 10, a classification module 11, a recovery prediction module 12 and a first determination module 13.
Specifically, the obtaining module 10 is configured to obtain feature data of the target undesirable asset;
the classification module 11 is configured to input the feature data into a preset classification model to obtain a classification result of the target undesirable asset;
the recovery rate prediction module 12 is configured to, when the classification result is not equal to a target classification result, input the feature data into a preset prediction model to obtain a recovery rate of the target undesirable asset, where the target classification result is related to an attribution attribute of the target undesirable asset;
the first determining module 13 is configured to determine a recovery price of the target bad asset according to the recovery rate.
According to the determining device for the recovery price of the poor assets, after the characteristic data of the target poor assets are obtained, the computer equipment inputs the characteristic data into the preset classification model to obtain the classification result of the target poor assets, when the classification result is not equal to the target classification result, the characteristic data are input into the preset prediction model to obtain the recovery rate of the target poor assets, and the recovery price of the target poor assets is determined according to the obtained recovery rate. In the whole pricing process of the target bad assets, the computer equipment can predict the recovery rate of the target bad assets based on the pre-trained classification model and the prediction model and determine the recovery price of the target bad assets based on the predicted recovery rate, namely, the whole pricing process is automatically completed by the computer equipment, and the asset security manager is not required to perform offline pricing based on own experience, so that the pricing efficiency of the bad assets is improved, and meanwhile, compared with manual pricing, the technical scheme also improves the accuracy of pricing results.
On the basis of the above embodiment, optionally, the apparatus further includes: a second determination module;
specifically, the second determining module is configured to determine the recovery price of the target bad asset based on the recovery rate corresponding to the target classification result when the classification result is equal to the target classification result.
Optionally, the classification model is obtained by performing model training based on a historical undesirable asset classification set and a preset first basic model, where the historical undesirable asset classification set includes feature data of historical undesirable assets and an actual classification result.
Optionally, the prediction model is obtained by performing model training based on a historical undesirable asset prediction set and a preset second basic model, where the historical undesirable asset prediction set includes characteristic data and an actual recovery rate of the historical undesirable asset.
Optionally, when the attribution attribute of the target bad asset is an enterprise type bad asset, the target classification result is zero recovery;
and when the attribution attribute of the target bad asset is personal bad assets, the target classification result is complete recovery.
On the basis of the above embodiment, optionally, the apparatus further includes: a preprocessing module;
specifically, the preprocessing module is configured to preprocess the feature data according to a data type of the feature data before the classification module 11 inputs the feature data into a preset classification model to obtain a classification result of the target undesirable asset.
On the basis of the foregoing embodiment, optionally, the first determining module 13 is specifically configured to obtain a credit balance of the target undesirable asset; and determining the recovery price of the target bad assets according to the recovery rate and the credit balance.
In one embodiment, a computer device is provided, a schematic structural diagram of which may be as shown in fig. 5. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used to store data during the determination of the recovery price for the undesirable asset. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of determining a recovery price for an undesirable asset.
Those skilled in the art will appreciate that the architecture shown in fig. 5 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, the computer device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the following steps when executing the computer program:
acquiring characteristic data of target bad assets;
inputting the characteristic data into a preset classification model to obtain a classification result of the target bad asset;
when the classification result is not equal to a target classification result, inputting the characteristic data into a preset prediction model to obtain the recovery rate of the target bad asset, wherein the target classification result is related to the attribution attribute of the target bad asset;
and determining the recovery price of the target bad assets according to the recovery rate.
In one embodiment, the processor, when executing the computer program, further performs the steps of: and when the classification result is equal to a target classification result, determining the recovery price of the target bad asset based on the recovery rate corresponding to the target classification result.
Optionally, the classification model is obtained by performing model training based on a historical undesirable asset classification set and a preset first basic model, where the historical undesirable asset classification set includes feature data of historical undesirable assets and an actual classification result.
Optionally, the prediction model is obtained by performing model training based on a historical undesirable asset prediction set and a preset second basic model, where the historical undesirable asset prediction set includes characteristic data and an actual recovery rate of the historical undesirable asset.
Optionally, when the attribution attribute of the target bad asset is an enterprise type bad asset, the target classification result is zero recovery; and when the attribution attribute of the target bad asset is personal bad assets, the target classification result is complete recovery.
In one embodiment, the processor, when executing the computer program, further performs the steps of: and preprocessing the characteristic data according to the data type of the characteristic data.
In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring the credit balance of the target bad assets; and determining the recovery price of the target bad assets according to the recovery rate and the credit balance.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring characteristic data of target bad assets;
inputting the characteristic data into a preset classification model to obtain a classification result of the target bad asset;
when the classification result is not equal to a target classification result, inputting the characteristic data into a preset prediction model to obtain the recovery rate of the target bad asset, wherein the target classification result is related to the attribution attribute of the target bad asset;
and determining the recovery price of the target bad assets according to the recovery rate.
In one embodiment, the computer program when executed by the processor further performs the steps of: and when the classification result is equal to a target classification result, determining the recovery price of the target bad asset based on the recovery rate corresponding to the target classification result.
Optionally, the classification model is obtained by performing model training based on a historical undesirable asset classification set and a preset first basic model, where the historical undesirable asset classification set includes feature data of historical undesirable assets and an actual classification result.
Optionally, the prediction model is obtained by performing model training based on a historical undesirable asset prediction set and a preset second basic model, where the historical undesirable asset prediction set includes characteristic data and an actual recovery rate of the historical undesirable asset.
Optionally, when the attribution attribute of the target bad asset is an enterprise type bad asset, the target classification result is zero recovery; and when the attribution attribute of the target bad asset is personal bad assets, the target classification result is complete recovery.
In one embodiment, the computer program when executed by the processor further performs the steps of: and preprocessing the characteristic data according to the data type of the characteristic data.
In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring the credit balance of the target bad assets; and determining the recovery price of the target bad assets according to the recovery rate and the credit balance.
The device for determining the recovery price of the undesirable asset, the computer device and the storage medium provided in the above embodiments may execute the method for determining the recovery price of the undesirable asset provided in any embodiment of the present application, and have corresponding functional modules and beneficial effects for executing the method. For technical details not described in detail in the above embodiments, reference may be made to a method for determining a recovery price of a bad asset provided in any of the embodiments of the present application.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method for determining a recovery price of an undesirable asset, comprising:
acquiring characteristic data of target bad assets;
inputting the characteristic data into a preset classification model to obtain a classification result of the target bad asset;
when the classification result is not equal to a target classification result, inputting the characteristic data into a preset prediction model to obtain the recovery rate of the target bad asset, wherein the target classification result is related to the attribution attribute of the target bad asset;
and determining the recovery price of the target bad assets according to the recovery rate.
2. The method of claim 1, further comprising:
and when the classification result is equal to a target classification result, determining the recovery price of the target bad asset based on the recovery rate corresponding to the target classification result.
3. The method according to claim 2, wherein the classification model is obtained by model training based on a historical undesirable asset classification set and a preset first basic model, wherein the historical undesirable asset classification set comprises characteristic data of historical undesirable assets and actual classification results.
4. The method of claim 2, wherein the predictive model is model trained based on a predictive set of historical undesirable assets and a pre-defined second base model, wherein the predictive set of historical undesirable assets includes characteristic data and actual recovery rates of historical undesirable assets.
5. The method according to any one of claims 1 to 4, wherein when the attribution attribute of the target bad asset is a business class bad asset, the target classification result is zero reclamation;
and when the attribution attribute of the target bad asset is personal bad assets, the target classification result is complete recovery.
6. The method according to any one of claims 1 to 4, wherein before the inputting the feature data into a preset classification model to obtain the classification result of the target bad asset, the method further comprises:
and preprocessing the characteristic data according to the data type of the characteristic data.
7. The method of any one of claims 1 to 4, wherein determining a recovery price for the target undesirable asset from the recovery rate comprises:
acquiring the credit balance of the target bad assets;
and determining the recovery price of the target bad assets according to the recovery rate and the credit balance.
8. An apparatus for determining a recovery price of an undesirable asset, comprising:
the acquisition module is used for acquiring the characteristic data of the target bad asset;
the classification module is used for inputting the characteristic data into a preset classification model to obtain a classification result of the target bad asset;
the recovery rate prediction module is used for inputting the characteristic data into a preset prediction model to obtain the recovery rate of the target bad asset when the classification result is not equal to a target classification result, wherein the target classification result is related to the attribution attribute of the target bad asset;
and the first determining module is used for determining the recovery price of the target bad asset according to the recovery rate.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010565628.1A CN111738762A (en) | 2020-06-19 | 2020-06-19 | Method, device, equipment and storage medium for determining recovery price of poor assets |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010565628.1A CN111738762A (en) | 2020-06-19 | 2020-06-19 | Method, device, equipment and storage medium for determining recovery price of poor assets |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111738762A true CN111738762A (en) | 2020-10-02 |
Family
ID=72650374
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010565628.1A Pending CN111738762A (en) | 2020-06-19 | 2020-06-19 | Method, device, equipment and storage medium for determining recovery price of poor assets |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111738762A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862586A (en) * | 2020-12-03 | 2021-05-28 | 浙江惠瀜网络科技有限公司 | Undesirable asset risk management system |
| CN114139490A (en) * | 2022-02-07 | 2022-03-04 | 建元和光(北京)科技有限公司 | A method, device and device for automatic data preprocessing |
| CN114219025A (en) * | 2021-12-14 | 2022-03-22 | 中国建设银行股份有限公司 | A kind of asset recovery rate classification method, device, equipment and storage medium |
| CN114331686A (en) * | 2021-12-30 | 2022-04-12 | 鲁信科技股份有限公司 | Method, device and medium for managing bad assets based on labels |
| CN115511187A (en) * | 2022-09-29 | 2022-12-23 | 中国建设银行股份有限公司 | Method, device, equipment, medium and computer program product for predicting return on assets |
| CN115841346A (en) * | 2023-02-23 | 2023-03-24 | 杭银消费金融股份有限公司 | Asset derating prediction method and system for business decisions |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106952155A (en) * | 2017-03-08 | 2017-07-14 | 深圳前海纵腾金融科技服务有限公司 | A kind of collection method and device based on credit scoring |
| CN109377333A (en) * | 2018-09-03 | 2019-02-22 | 平安科技（深圳）有限公司 | Electronic device determines method and storage medium based on the collection person of disaggregated model |
| CN109559221A (en) * | 2018-11-20 | 2019-04-02 | 中国银行股份有限公司 | Collection method, apparatus and storage medium based on user data |
| US20190155941A1 (en) * | 2017-11-21 | 2019-05-23 | International Business Machines Corporation | Generating asset level classifications using machine learning |
| CN110020939A (en) * | 2019-03-01 | 2019-07-16 | 平安科技（深圳）有限公司 | Establish device, method and the storage medium of loss given default prediction model |
-
2020
- 2020-06-19 CN CN202010565628.1A patent/CN111738762A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106952155A (en) * | 2017-03-08 | 2017-07-14 | 深圳前海纵腾金融科技服务有限公司 | A kind of collection method and device based on credit scoring |
| US20190155941A1 (en) * | 2017-11-21 | 2019-05-23 | International Business Machines Corporation | Generating asset level classifications using machine learning |
| CN109377333A (en) * | 2018-09-03 | 2019-02-22 | 平安科技（深圳）有限公司 | Electronic device determines method and storage medium based on the collection person of disaggregated model |
| CN109559221A (en) * | 2018-11-20 | 2019-04-02 | 中国银行股份有限公司 | Collection method, apparatus and storage medium based on user data |
| CN110020939A (en) * | 2019-03-01 | 2019-07-16 | 平安科技（深圳）有限公司 | Establish device, method and the storage medium of loss given default prediction model |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862586A (en) * | 2020-12-03 | 2021-05-28 | 浙江惠瀜网络科技有限公司 | Undesirable asset risk management system |
| CN114219025A (en) * | 2021-12-14 | 2022-03-22 | 中国建设银行股份有限公司 | A kind of asset recovery rate classification method, device, equipment and storage medium |
| CN114331686A (en) * | 2021-12-30 | 2022-04-12 | 鲁信科技股份有限公司 | Method, device and medium for managing bad assets based on labels |
| CN114139490A (en) * | 2022-02-07 | 2022-03-04 | 建元和光(北京)科技有限公司 | A method, device and device for automatic data preprocessing |
| CN114139490B (en) * | 2022-02-07 | 2022-08-02 | 建元和光(北京)科技有限公司 | A method, device and device for automatic data preprocessing |
| CN115511187A (en) * | 2022-09-29 | 2022-12-23 | 中国建设银行股份有限公司 | Method, device, equipment, medium and computer program product for predicting return on assets |
| CN115841346A (en) * | 2023-02-23 | 2023-03-24 | 杭银消费金融股份有限公司 | Asset derating prediction method and system for business decisions |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111738762A (en) | Method, device, equipment and storage medium for determining recovery price of poor assets | |
| Rao et al. | Design of comprehensive evaluation index system for P2P credit risk of “three rural” borrowers | |
| Rao et al. | Credit risk assessment mechanism of personal auto loan based on PSO-XGBoost Model | |
| Kwak et al. | Bankruptcy prediction for Korean firms after the 1997 financial crisis: using a multiple criteria linear programming data mining approach | |
| CN113327164A (en) | Risk control method and device for futures trading and computer equipment | |
| Chaudhari et al. | Synergizing Generative AI and Machine Learning for Financial Credit Risk Forecasting and Code Auditing | |
| CN114202399A (en) | Intelligent approval method and related device | |
| CN115860889A (en) | Financial loan big data management method and system based on artificial intelligence | |
| CN114020798A (en) | Data classification method, device, equipment and storage medium | |
| CN117575773A (en) | Method, device, computer equipment and storage medium for determining service data | |
| Sharma et al. | A Logistic Regression Based Credit Risk Assessment Using WoE Bining and Enhanced Feature Engineering Approach ANOVA and Chi-Square | |
| Lee et al. | Application of machine learning in credit risk scorecard | |
| Yang | A random forest approach to appraise personal credit risk of internet loans | |
| Septama et al. | A Comparative Analysis of Machine Learning Algorithms for Credit Risk Scoring using Chi-Square Feature Selection | |
| CN117764692A (en) | Method for predicting credit risk default probability | |
| CN114581248A (en) | Client investment risk preference prediction method and device | |
| CN113762415A (en) | A method and system for intelligent matching of auto financial products based on neural network | |
| Kaisar et al. | Explainable Machine Learning Models for Credit Risk Analysis: A Survey | |
| Nanda et al. | Big Data Analytics Function in Financial Institutions for Risk Management Practices | |
| CN115757971B (en) | Business object recommendation method, device, computer equipment and storage medium | |
| Nayak et al. | Predicting Stock Market Price Using Machine Learning Techniques | |
| Saranya | Utilizing AI and Machine Learning in Financial Analysis | |
| Rashad et al. | The Prediction of Non-Performing Loans Using Artificial Intelligence-A literature Review | |
| Xu et al. | Labor Market Employment Rate Prediction and Optimization Based on Deep Neural Networks | |
| CN118333739A (en) | Method for constructing retail credit risk prediction model and retail credit business Scoremult model |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20220915 Address after: 25 Financial Street, Xicheng District, Beijing 100033 Applicant after: CHINA CONSTRUCTION BANK Corp. Address before: 25 Financial Street, Xicheng District, Beijing 100033 Applicant before: CHINA CONSTRUCTION BANK Corp. Applicant before: Jianxin Financial Science and Technology Co.,Ltd. |