Credit investigation data processing method and device, equipment, storage medium and product

Abstract

The invention discloses a credit investigation data processing method and device, equipment, a storage medium and a product. The invention relates to the technical field of big data. The method comprises the steps of obtaining credit investigation text data of a target object; inputting the credit investigation text data of the target object into a pre-constructed expert experience score card model to obtain expert experience risk prediction data of the target object; inputting the credit investigation text data of the target object into a pre-constructed integrated learning score card model to obtain integrated learning risk prediction data of the target object; and determining target risk prediction data of the target object based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object. According to the technical scheme, the technical effect of improving the risk prediction accuracy can be achieved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of big data technology, and in particular, to a credit information data processing method, device, equipment, storage medium and product. Background Art

[0002] With the development of big data technology, it has been increasingly widely used in the field of financial risk control.

[0003] At present, credit risk assessment usually selects indicators and sets scores corresponding to the indicators based on expert experience, and then constructs a scorecard to conduct credit risk assessment on users based on the scorecard.

[0004] In the process of implementing the present invention, it is found that there are at least the following technical problems in the prior art: the existing risk prediction scheme has the problem of inaccurate scoring. Summary of the invention

[0005] Embodiments of the present invention provide a credit data processing method, device, equipment, storage medium and product to solve the problem of inaccurate risk prediction.

[0006] In a first aspect, an embodiment of the present invention provides a credit information data processing method, comprising:

[0007] Obtain the target object's credit text data;

[0008] Inputting the credit report text data of the target object into a pre-built expert experience scorecard model to obtain the expert experience risk prediction data of the target object;

[0009] Inputting the credit report text data of the target object into a pre-built integrated learning scorecard model to obtain the integrated learning risk prediction data of the target object;

[0010] Based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object, target risk prediction data of the target object is determined.

[0011] In a second aspect, an embodiment of the present invention further provides a credit information data processing device, the device comprising:

[0012] A credit information data acquisition module is used to acquire the credit information text data of the target object;

[0013] An expert experience scoring module is used to input the credit report text data of the target object into a pre-built expert experience scoring card model to obtain the expert experience risk prediction data of the target object;

[0014] An integrated learning scoring module, used to input the credit report text data of the target object into a pre-built integrated learning scoring card model to obtain the integrated learning risk prediction data of the target object;

[0015] The target risk prediction data determination module is used to determine the target risk prediction data of the target object based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object.

[0016] In a third aspect, an embodiment of the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a credit data processing method as described in any one of the embodiments of the present invention is implemented.

[0017] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a credit data processing method as described in any one of the embodiments of the present invention.

[0018] In a fifth aspect, an embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements a credit data processing method as described in any one of the embodiments of the present invention.

[0019] In the embodiment of the present invention, the target object's credit text data is obtained, and then the target object's credit text data is input into a pre-built expert experience scorecard model to obtain the target object's expert experience risk prediction data; in addition, the target object's credit text data is input into a pre-built integrated learning scorecard model to obtain the target object's integrated learning risk prediction data, and then based on the target object's expert experience risk prediction data and the target object's integrated learning risk prediction data, the target object's target risk prediction data is determined. The technical solution of the present invention comprehensively determines the risk prediction data by combining the results of the expert experience scorecard and the integrated learning scorecard, takes into account the quality differences of different risk prediction methods, and improves the accuracy of the fused risk prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A flowchart of a credit information data processing method provided by an embodiment of the present invention;

[0022] Figure 2 A flowchart of a credit information data processing method provided by an embodiment of the present invention;

[0023] Figure 3 A flowchart of a credit information data processing method provided by an embodiment of the present invention;

[0024] Figure 4 A flowchart of a credit information data processing method provided by an embodiment of the present invention;

[0025] Figure 5 A schematic diagram of the structure of a credit information data processing device provided by an embodiment of the present invention;

[0026] Figure 6 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0028] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance. The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0029] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0030] Figure 1 This is a flow chart of a credit data processing method provided by an embodiment of the present invention. This embodiment is applicable to the case of credit risk assessment through a scoring card. The method can be executed by a credit data processing device. The credit data processing device can be implemented in the form of hardware and / or software. The credit data processing device can be configured in electronic devices such as terminals and / or servers. Figure 1 As shown, the method includes:

[0031] S110, obtaining the credit information text data of the target object.

[0032] In the embodiment of the present invention, the target object may be a borrower or company of a financial institution, etc., and may be an object to be subject to credit risk assessment. Credit investigation text data refers to the credit behavior record of the target object, which may include but is not limited to risk assessment indicators such as the target object's age, monthly income, overdue interest, debt-to-asset ratio, and net profit growth rate.

[0033] Exemplarily, the credit data of the subject to be subject to credit risk assessment can be retrieved from a preset storage path of the electronic device, or the credit data of the subject to be subject to credit risk assessment sent by other devices that are communicatively connected to the electronic device can be received, without specific limitation here.

[0034] S120: Input the credit report text data of the target object into a pre-built expert experience scoring card model to obtain the expert experience risk prediction data of the target object.

[0035] In the embodiment of the present invention, the expert experience scoring card model refers to a scoring card developed based on expert experience, and the expert experience scoring card model may include multiple risk assessment indicators and scores corresponding to each risk assessment indicator. Specifically, the expert experience scoring card model may be used to score each risk assessment indicator value in the credit report text data of the target object, thereby calculating the expert experience score based on the scoring results of each risk assessment indicator value, that is, obtaining the expert experience risk prediction data. In other words, the risk prediction data may be a score.

[0036] Exemplarily, the risk assessment indicators in the expert experience scorecard model may include, but are not limited to, the company's years of establishment and registered capital, among which, if the company's years of establishment are less than two years, 1 point is corresponding, if the company's years of establishment are two to five years, 2 points are corresponding, if the company's years of establishment are five to eight years, 3 points are corresponding, and if the company's years of establishment are greater than eight years, 4 points are corresponding; if the registered capital is less than one million, 0 points are corresponding, if the registered capital is one to three million, 1 point is corresponding, if the registered capital is three to five million, 2 points are corresponding, and if the registered capital is greater than five million, 3 points are corresponding. If the company's years of establishment in the target object's credit report text data are ten years and the registered capital is twenty million, the expert experience score can be: 4+3=7 points.

[0037] S130: Input the credit report text data of the target object into a pre-built integrated learning scorecard model to obtain the integrated learning risk prediction data of the target object.

[0038] In an embodiment of the present invention, the ensemble learning scorecard model refers to a scorecard developed according to an ensemble learning algorithm, wherein the ensemble learning algorithm may be a gradient boosting decision tree (Light Gradient Boosting Machine, LightGBM) or other ensemble learning algorithms, which are not specifically limited herein.

[0039] Specifically, the target object's credit text data can be input into a pre-built integrated learning scoring card model. The integrated learning scoring card model performs risk prediction based on the target object's credit text data, thereby predicting the target object's integrated learning score, that is, obtaining integrated learning risk prediction data.

[0040] S140. Determine target risk prediction data of the target object based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object.

[0041] In an embodiment of the present invention, the expert experience risk prediction data and the integrated learning risk prediction data may be directly added or weighted fused to obtain target risk prediction data of the target object.

[0042] In the embodiment of the present invention, the target object's credit text data is obtained, and then the target object's credit text data is input into a pre-built expert experience scorecard model to obtain the target object's expert experience risk prediction data; in addition, the target object's credit text data is input into a pre-built integrated learning scorecard model to obtain the target object's integrated learning risk prediction data, and then based on the target object's expert experience risk prediction data and the target object's integrated learning risk prediction data, the target object's target risk prediction data is determined. The technical solution of the present invention comprehensively determines the risk prediction data by combining the results of the expert experience scorecard and the integrated learning scorecard, takes into account the quality differences of different risk prediction methods, and improves the accuracy of the fused risk prediction.

[0043] Figure 2A flow chart of a credit data processing method provided in an embodiment of the present invention, the method of this embodiment can be combined with the various optional schemes in the credit data processing method provided in the above embodiments. The credit data processing method provided in this embodiment has been further optimized. Optionally, the steps of constructing the expert experience scorecard model include: obtaining historical incomplete credit text data of multiple objects, wherein the data missing rate of the historical incomplete credit text data is greater than a preset missing rate threshold; performing qualitative comparative analysis based on the historical incomplete credit text data of the multiple objects to obtain target risk assessment indicators and evaluation values ​​corresponding to the target risk assessment indicators; constructing an expert experience scorecard model based on the target risk assessment indicators and the evaluation values ​​corresponding to the target risk assessment indicators.

[0044] like Figure 2 As shown, the method includes:

[0045] S210. Obtain incomplete historical credit investigation text data of multiple objects, wherein a data missing rate of the incomplete historical credit investigation text data is greater than a preset missing rate threshold.

[0046] In an embodiment of the present invention, the historical incomplete credit text data refers to historical credit data with serious data missing, which can be used to construct an expert experience scorecard model.

[0047] For example, for multiple customers with outstanding loans in a financial institution, the credit data of the above customers with a missing rate greater than 40% in the past five years can be used as historical incomplete credit text data, where the data missing rate is calculated as: data missing rate = number of missing values / total data.

[0048] S220: Performing a qualitative comparative analysis based on the historical incomplete credit investigation text data of the multiple objects to obtain a target risk assessment indicator and an evaluation value corresponding to the target risk assessment indicator.

[0049] In the embodiment of the present invention, Qualitative Comparative Analysis (QCA) refers to a data processing method that combines the advantages of qualitative and quantitative analysis methods. The target risk evaluation index refers to the risk evaluation index of the final selected score card obtained by qualitative comparative analysis screening. The evaluation value can be a score.

[0050] Specifically, the historical non-full credit investigation text data of multiple objects can be subjected to qualitative comparative analysis such as formulating a long list of indicators, single factor analysis, multi-factor analysis, and factor score segmentation, so as to obtain the target risk assessment indicators and the evaluation values ​​corresponding to the target risk assessment indicators. Specifically, formulating a long list of indicators refers to screening out risk assessment indicators that can fully cover the risk characteristics of customers from the historical non-full credit investigation text data. Single factor analysis refers to screening out the risk assessment indicators with the highest number of votes based on the voting results of the questionnaires issued on the basis of formulating a long list of indicators. Multi-factor analysis refers to determining the target risk assessment indicators and the corresponding weights of the target risk assessment indicators that are finally selected for the scoring card through the hierarchical analysis method based on the risk assessment indicators with the highest number of votes. Factor score segmentation refers to determining the evaluation value of the target risk assessment indicator based on statistical methods and expert experience. For example, when the target risk assessment indicator is a continuous variable, such as age, it can be divided into different intervals according to the age group, and a corresponding score is assigned to each age group interval; when the target risk assessment indicator is a discrete variable, such as education, it can be assigned a corresponding score according to the weight corresponding to the education level.

[0051] Optionally, a qualitative comparative analysis is performed based on historical incomplete credit text data of multiple objects to obtain a target risk assessment indicator and an evaluation value corresponding to the target risk assessment indicator, including: obtaining a first risk assessment indicator in response to a user's indicator selection operation on the historical incomplete credit text data of multiple objects; determining a second risk assessment indicator based on the voting result in response to the user's voting operation on the first risk assessment indicator; performing a hierarchical analysis on the second risk assessment indicator to obtain a target risk assessment indicator and a weight corresponding to the target risk assessment indicator; and determining an evaluation value corresponding to the target risk assessment indicator based on the weight corresponding to the target risk assessment indicator.

[0052] Exemplarily, a user can select a first risk assessment indicator that can comprehensively cover the risk characteristics of customers from historical incomplete credit text data of multiple objects, and then set up a questionnaire based on the first risk assessment indicator, and screen out a second risk assessment indicator with a high number of votes based on the voting results of the questionnaire, and then perform a hierarchical analysis on the second risk assessment indicator to obtain a target risk assessment indicator and a weight corresponding to the target risk assessment indicator. Then, the user can refer to the weight corresponding to the target risk assessment indicator and reasonably set the evaluation value corresponding to the target risk assessment indicator.

[0053] Among them, the first risk assessment indicator and the second risk assessment indicator are both selected candidate risk assessment indicators for subsequent screening.

[0054] Optionally, a hierarchical analysis is performed on the second risk assessment indicator to obtain a target risk assessment indicator and a weight corresponding to the target risk assessment indicator, including: hierarchical division of the second risk assessment indicator to obtain risk assessment indicators in the target layer, risk assessment indicators in the criterion layer, and risk assessment indicators in the scheme layer; a judgment matrix is ​​determined based on the risk assessment indicators in the target layer, the risk assessment indicators in the criterion layer, and the risk assessment indicators in the scheme layer; and a target risk assessment indicator and a weight corresponding to the target risk assessment indicator are determined based on the judgment matrix.

[0055] Among them, the target layer, criterion layer and solution layer are the hierarchical structures obtained by hierarchical division. The judgment matrix refers to the judgment given on the relative importance of each indicator in each level, and the judgment is expressed in numerical values ​​and written in matrix form.

[0056] Exemplarily, the risk assessment indicator in the target layer may be "assessing the credit risk of the loan applicant". The risk assessment indicators in the criterion layer may include "the repayment ability of the loan applicant", "the repayment willingness of the loan applicant" and "the credit record of the loan applicant", etc. The risk assessment indicators in the scheme layer may include "approving the loan", "rejecting the loan" and "adjusting the loan amount", etc. Further, the risk assessment indicators in the target layer, the risk assessment indicators in the criterion layer and the risk assessment indicators in the scheme layer may be compared in pairs to obtain a judgment matrix, and then the target risk assessment indicators and the weights corresponding to the target risk assessment indicators may be obtained by solving the judgment matrix.

[0057] S230: construct an expert experience scorecard model based on the target risk assessment indicator and the evaluation value corresponding to the target risk assessment indicator.

[0058] Exemplarily, the target risk evaluation index and the evaluation value corresponding to the target risk evaluation index may be input into a scorecard template, thereby obtaining an expert experience scorecard model.

[0059] S240: Obtain credit information text data of the target object.

[0060] S250: Input the credit report text data of the target object into a pre-built expert experience scorecard model to obtain the expert experience risk prediction data of the target object.

[0061] S260: Input the credit report text data of the target object into a pre-built integrated learning scorecard model to obtain the integrated learning risk prediction data of the target object.

[0062] S270: Determine target risk prediction data of the target object based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object.

[0063] In the embodiment of the present invention, by performing qualitative comparative analysis on historical incomplete credit report text data, the development and construction of an expert experience scorecard model is achieved, providing an accurate expert experience scorecard model for subsequent risk prediction.

[0064] Figure 3 A flow chart of a credit data processing method provided in an embodiment of the present invention, the method of this embodiment can be combined with the various optional schemes in the credit data processing method provided in the above embodiments. The credit data processing method provided in this embodiment has been further optimized. Optionally, the steps of constructing the integrated learning scorecard model include: obtaining historical full credit text data of multiple objects, wherein the data missing rate of the historical full credit text data is less than a preset missing rate threshold; performing chi-square binning based on the historical full credit text data of the multiple objects to obtain binning results; performing evidence weight conversion on the binning results to obtain the evidence weight value corresponding to the binning results; based on the evidence weight value corresponding to the binning results, training the initial machine learning model to obtain the integrated learning scorecard model.

[0065] like Figure 3 As shown, the method includes:

[0066] S310. Obtaining historical full credit text data of multiple objects, wherein a data missing rate of the historical full credit text data is less than a preset missing rate threshold.

[0067] In the embodiment of the present invention, the historical full credit text data refers to the historical credit data with complete data or minor data missing, which can be used to construct an integrated learning scorecard model.

[0068] For example, for multiple customers with outstanding loans in a financial institution, the credit data of the above customers with a missing rate of less than 40% in the past five years can be used as the historical full credit text data.

[0069] S320: Perform chi-square binning based on the historical full credit investigation text data of the multiple objects to obtain binning results.

[0070] In an embodiment of the present invention, the historical full amount of credit report text data can be divided into several intervals (also called "buckets" or "boxes") through chi-square binning, and each interval is assigned a discretized value for better modeling and analysis.

[0071] S330: Perform evidence weight conversion on the binning result to obtain the evidence weight value corresponding to the binning result.

[0072] In the embodiment of the present invention, by performing weight of evidence (WOE) conversion on the binning result, the weight of evidence value corresponding to the binning result can be obtained.

[0073] S340. Based on the evidence weight values ​​corresponding to the binning results, the initial machine learning model is trained to obtain an integrated learning scorecard model.

[0074] In an embodiment of the present invention, the initial machine learning model may be a gradient boosting decision tree (Light Gradient Boosting Machine, LightGBM) or other integrated learning models.

[0075] Specifically, the gradient boosting decision tree can be trained based on the weight of evidence corresponding to the binning results, thereby obtaining an integrated learning scorecard model. Specifically, the weight of evidence corresponding to the binning results can be input into the gradient boosting decision tree to obtain decision tree information, wherein the decision tree information includes the depth of the tree, the number of leaf nodes, and the output value of the leaf nodes; the decision tree information is then input into a loss function with a regular term to obtain a model loss value, and then the model parameters of the gradient boosting decision tree are adjusted based on the model loss value until the training stop condition is met to obtain an integrated learning scorecard model.

[0076] It should be noted that by adding a regularization term to the loss function, the complexity of the gradient boosting decision tree model can be controlled to reduce the risk of overfitting.

[0077] S350: Obtain the credit information text data of the target object.

[0078] S360: Input the credit report text data of the target object into a pre-built expert experience scorecard model to obtain the expert experience risk prediction data of the target object.

[0079] S370: Input the credit report text data of the target object into a pre-built integrated learning scorecard model to obtain the integrated learning risk prediction data of the target object.

[0080] S380: Determine target risk prediction data of the target object based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object.

[0081] Optionally, after performing ensemble learning training on the initial machine learning model based on the evidence weight value corresponding to the binning result to obtain the ensemble learning scoring card model, it also includes: performing parameter tuning on the ensemble learning scoring card model to obtain the ensemble learning scoring card model after parameter tuning; correspondingly, inputting the credit text data of the target object into the pre-built ensemble learning scoring card model to obtain the ensemble learning risk prediction data of the target object, including: inputting the credit data of the target object into the ensemble learning scoring card model after parameter tuning to obtain the ensemble learning risk prediction data of the target object.

[0082] Exemplarily, parameters such as learning rate, number of iterations, maximum depth of the tree, proportion of data used in each iteration, and proportion of features randomly selected in each iteration can be adjusted to complete parameter tuning of the integrated learning scorecard model.

[0083] It should be noted that by tuning the parameters of the ensemble learning scorecard model, the prediction accuracy of the ensemble learning scorecard model can be improved, thereby improving the accuracy of the ensemble learning score.

[0084] In an embodiment of the present invention, the historical full amount of credit report text data is subjected to chi-square binning to obtain the binning results, and then the binning results are converted into evidence weights to obtain the evidence weight values ​​corresponding to the binning results. Then, based on the evidence weight values ​​corresponding to the binning results, the initial machine learning model is trained to obtain an integrated learning scorecard model, thereby realizing the development and construction of the integrated learning scorecard model and providing an accurate integrated learning scorecard model for subsequent risk prediction.

[0085] Figure 4 A flowchart of a credit data processing method provided in an embodiment of the present invention, the method of this embodiment can be combined with each optional scheme in the credit data processing method provided in the above embodiment. The credit data processing method provided in this embodiment has been further optimized. Optionally, the target risk prediction data of the target object is determined based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object, including: determining the weight information of the expert experience risk prediction data and the weight information corresponding to the integrated learning risk prediction data; determining the first risk prediction data based on the expert experience risk prediction data of the target object and the weight information of the expert experience risk prediction data; determining the second risk prediction data based on the integrated learning risk prediction data of the target object and the weight information corresponding to the integrated learning risk prediction data; determining the target risk prediction data of the target object based on the first risk prediction data and the second risk prediction data.

[0086] like Figure 4 As shown, the method includes:

[0087] S410: Obtain credit information text data of the target object.

[0088] S420: Input the credit report text data of the target object into a pre-built expert experience scorecard model to obtain the expert experience risk prediction data of the target object.

[0089] S430: Input the credit report text data of the target object into a pre-built integrated learning scorecard model to obtain the integrated learning risk prediction data of the target object.

[0090] S440, determining weight information of the expert experience risk prediction data and weight information corresponding to the integrated learning risk prediction data.

[0091] In some embodiments, the user can customize the input of the weight information of the expert experience risk prediction data and the weight information corresponding to the integrated learning risk prediction data through the interactive interface. In some embodiments, the weight information of the expert experience risk prediction data and the weight information of the integrated learning risk prediction data can also be adjusted according to the data missing rate or the accuracy of the scorecard model, for example, a higher weight is assigned to the risk prediction data with a low data missing rate.

[0092] S450: Determine first risk prediction data based on the expert experience risk prediction data of the target object and weight information of the expert experience risk prediction data.

[0093] Exemplarily, the first risk prediction data=expert experience risk prediction data×weight information of the expert experience risk prediction data.

[0094] S460. Determine second risk prediction data based on the integrated learning risk prediction data of the target object and weight information corresponding to the integrated learning risk prediction data.

[0095] Exemplarily, the second risk prediction data=ensemble learning risk prediction data×weight information corresponding to the ensemble learning risk prediction data.

[0096] S470. Determine target risk prediction data of the target object based on the first risk prediction data and the second risk prediction data.

[0097] Exemplarily, target risk prediction data=first risk prediction data+second risk prediction data.

[0098] It should be noted that by weighted fusion of expert experience risk prediction data and ensemble learning risk prediction data, the quality differences of different credit data processing methods are taken into account, effectively improving the accuracy of risk prediction.

[0099] Optionally, determining the weight information of the expert experience risk prediction data and the weight information corresponding to the integrated learning risk prediction data includes: obtaining the data missing rate of historical incomplete credit text data; determining the weight information of the expert experience risk prediction data based on the data missing rate of the historical incomplete credit text data; obtaining the data missing rate of historical complete credit text data; and determining the weight information corresponding to the integrated learning risk prediction data based on the data missing rate of the historical complete credit text data.

[0100] In the embodiment of the present invention, the weight information may be determined according to a calculation formula or a mapping relationship between the data missing rate and the weight.

[0101] For example, if the data missing rate of historical incomplete credit text data is 70%, the weight information of expert experience risk prediction data can be: 1-70%=0.3; if the data missing rate of historical full credit text data is 30%, the corresponding weight information of integrated learning risk prediction data can be 1-30%=0.7.

[0102] Optionally, determining the weight information of the expert experience risk prediction data and the weight information corresponding to the integrated learning risk prediction data includes: obtaining the prediction accuracy of the expert experience scoring card model; determining the weight information of the expert experience risk prediction data based on the prediction accuracy of the expert experience scoring card model; obtaining the prediction accuracy of the integrated learning scoring card model; and determining the weight information corresponding to the integrated learning risk prediction data based on the prediction accuracy of the integrated learning scoring card model.

[0103] In the embodiment of the present invention, the weight information may be determined according to a calculation formula or a mapping relationship between the prediction accuracy and the weight.

[0104] For example, if the prediction accuracy of the expert experience scorecard model is 60%, the weight information of the expert experience score can be 0.6, and if the prediction accuracy of the integrated learning scorecard model is 40%, the corresponding weight information of the integrated learning score can be 0.4.

[0105] In the embodiment of the present invention, by weighted fusion of expert experience risk prediction data and ensemble learning risk prediction data, the quality differences of different credit data processing methods are taken into account, thereby improving the accuracy of the fused risk prediction.

[0106] Figure 5 A schematic diagram of the structure of a credit data processing device provided by an embodiment of the present invention. Figure 5 As shown, the device comprises:

[0107] The credit information data acquisition module 510 is used to acquire the credit information text data of the target object;

[0108] The expert experience scoring module 520 is used to input the credit report text data of the target object into the pre-built expert experience scoring card model to obtain the expert experience risk prediction data of the target object;

[0109] The integrated learning scoring module 530 is used to input the credit report text data of the target object into the pre-built integrated learning scoring card model to obtain the integrated learning risk prediction data of the target object;

[0110] The target risk prediction data determination module 540 is used to determine the target risk prediction data of the target object based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object.

[0111] In the embodiment of the present invention, the target object's credit text data is obtained, and then the target object's credit text data is input into a pre-built expert experience scorecard model to obtain the target object's expert experience risk prediction data; in addition, the target object's credit text data is input into a pre-built integrated learning scorecard model to obtain the target object's integrated learning risk prediction data, and then based on the target object's expert experience risk prediction data and the target object's integrated learning risk prediction data, the target object's target risk prediction data is determined. The technical solution of the present invention comprehensively determines the risk prediction data by combining the results of the expert experience scorecard and the integrated learning scorecard, takes into account the quality differences of different risk prediction methods, and improves the accuracy of the fused risk prediction.

[0112] In some optional implementations, the credit information data processing device further includes:

[0113] A historical incomplete credit investigation data acquisition module, used to acquire historical incomplete credit investigation text data of multiple objects, wherein the data missing rate of the historical incomplete credit investigation text data is greater than a preset missing rate threshold;

[0114] A qualitative comparative analysis module, used to perform qualitative comparative analysis based on the historical incomplete credit investigation text data of the multiple objects, and obtain a target risk evaluation index and an evaluation value corresponding to the target risk evaluation index;

[0115] The expert experience scorecard model construction module is used to construct an expert experience scorecard model based on the target risk evaluation indicator and the evaluation value corresponding to the target risk evaluation indicator.

[0116] In some optional embodiments, the qualitative comparative analysis module includes:

[0117] An indicator selection unit, configured to obtain a first risk assessment indicator in response to a user's indicator selection operation on the historical partial credit investigation text data of the plurality of objects;

[0118] An indicator voting unit, configured to respond to a user's voting operation on the first risk assessment indicator and determine a second risk assessment indicator based on a voting result;

[0119] The hierarchical analysis unit is used to perform hierarchical analysis on the second risk assessment indicator to obtain a target risk assessment indicator and an evaluation value corresponding to the target risk assessment indicator.

[0120] In some optional implementations, the hierarchical analysis unit may also be used to:

[0121] The second risk assessment index is hierarchically divided to obtain risk assessment indexes in a target layer, risk assessment indexes in a criterion layer, and risk assessment indexes in a solution layer;

[0122] Determine a judgment matrix based on the risk assessment indicators in the target layer, the risk assessment indicators in the criterion layer, and the risk assessment indicators in the solution layer;

[0123] A target risk assessment indicator and an assessment value corresponding to the target risk assessment indicator are determined based on the judgment matrix.

[0124] In some optional implementations, the credit information data processing device further includes:

[0125] A historical full credit text data acquisition module, used to acquire historical full credit text data of multiple objects, wherein the data missing rate of the historical full credit text data is less than a preset missing rate threshold;

[0126] A chi-square binning module, used to perform chi-square binning based on the historical full amount of credit investigation text data of the multiple objects to obtain binning results;

[0127] An evidence weight conversion module is used to perform evidence weight conversion on the binning results to obtain the evidence weight value corresponding to the binning results;

[0128] The model training module is used to train the initial machine learning model based on the evidence weight value corresponding to the binning result to obtain an integrated learning scorecard model.

[0129] In some optional implementations, the model training module includes:

[0130] The gradient boosting decision tree training unit is used to train the gradient boosting decision tree based on the evidence weight value corresponding to the binning result to obtain an integrated learning scorecard model.

[0131] In some optional implementations, the gradient boosting decision tree training unit may be specifically used for:

[0132] Inputting the evidence weight value corresponding to the binning result into the gradient boosting decision tree to obtain decision tree information, wherein the decision tree information includes the depth of the tree, the number of leaf nodes and the output value of the leaf nodes;

[0133] The decision tree information is input into a loss function with a regularization term to obtain a model loss value, and the model parameters of the gradient boosting decision tree are adjusted based on the model loss value until the training stop condition is met to obtain an integrated learning scorecard model.

[0134] In some optional implementations, the credit information data processing device further includes:

[0135] A parameter tuning module, used to perform parameter tuning on the integrated learning scorecard model to obtain the integrated learning scorecard model after parameter tuning;

[0136] Accordingly, the integrated learning scoring module 530 is also used for:

[0137] The credit report text data of the target object is input into the integrated learning scorecard model after parameter tuning to obtain the integrated learning risk prediction data of the target object.

[0138] In some optional implementations, the target risk prediction data determination module 540 includes:

[0139] A weight information determination unit, used to determine the weight information of the expert experience risk prediction data and the weight information corresponding to the ensemble learning risk prediction data;

[0140] A first risk prediction determination unit, configured to determine first risk prediction data based on the expert experience risk prediction data of the target object and weight information of the expert experience risk prediction data;

[0141] A second risk prediction determination unit, configured to determine second risk prediction data based on the integrated learning risk prediction data of the target object and weight information corresponding to the integrated learning risk prediction data;

[0142] A target risk prediction determination unit is used to determine the target risk prediction data of the target object based on the first risk prediction data and the second risk prediction data.

[0143] In some optional implementations, the weight information determination unit may be specifically configured to:

[0144] Obtain the data missing rate of historical incomplete credit report text data;

[0145] Determining the weight information of the expert experience risk prediction data based on the data missing rate of the historical incomplete credit investigation text data;

[0146] Obtain the data missing rate of the entire historical credit report text data;

[0147] The weight information corresponding to the integrated learning risk prediction data is determined based on the data missing rate of the historical full credit text data.

[0148] In some optional implementations, the weight information determination unit may be specifically configured to:

[0149] Obtain the prediction accuracy of the expert experience scorecard model;

[0150] Determining weight information of the expert experience risk prediction data based on the prediction accuracy of the expert experience scorecard model;

[0151] Get the prediction accuracy of the ensemble learning scorecard model;

[0152] The weight information corresponding to the ensemble learning risk prediction data is determined based on the prediction accuracy of the ensemble learning scorecard model.

[0153] The credit data processing device provided in the embodiment of the present invention can execute the credit data processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0154] Figure 6 This is a schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure. Figure 6 , which shows an electronic device (eg, Figure 6 The terminal device in the embodiment of the present disclosure may include but is not limited to mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0155] like Figure 6As shown, the electronic device 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An edit / output (I / O) interface 505 is also connected to the bus 504.

[0156] Typically, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 500 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 6 The electronic device 500 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0157] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0158] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0159] The electronic device provided in the embodiment of the present disclosure and the credit data processing method provided in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0160] The embodiments of the present disclosure provide a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the credit data processing method provided in the above embodiments is implemented.

[0161] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0162] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0163] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0164] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device:

[0165] Obtain the target object's credit text data;

[0166] Inputting the credit report text data of the target object into a pre-built expert experience scorecard model to obtain the expert experience risk prediction data of the target object;

[0167] Inputting the credit report text data of the target object into a pre-built integrated learning scorecard model to obtain the integrated learning risk prediction data of the target object;

[0168] Based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object, target risk prediction data of the target object is determined.

[0169] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0170] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0171] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware, wherein the name of a unit does not, in some cases, limit the unit itself.

[0172] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0173] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0174] An embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the credit data processing method provided in any embodiment of the present application.

[0175] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0176] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A credit investigation data processing method, characterized in that: include: Obtain the target object's credit text data; Inputting the credit report text data of the target object into a pre-built expert experience scorecard model to obtain the expert experience risk prediction data of the target object; Inputting the credit report text data of the target object into a pre-built integrated learning scorecard model to obtain the integrated learning risk prediction data of the target object; Based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object, target risk prediction data of the target object is determined.

2. The method according to claim 1, characterized in that The steps of constructing the expert experience scorecard model include: Acquiring incomplete historical credit investigation text data of multiple objects, wherein a data missing rate of the incomplete historical credit investigation text data is greater than a preset missing rate threshold; Performing qualitative comparative analysis based on the historical partial credit investigation text data of the multiple objects to obtain target risk evaluation indicators and evaluation values ​​corresponding to the target risk evaluation indicators; An expert experience scorecard model is constructed based on the target risk assessment indicator and the evaluation value corresponding to the target risk assessment indicator.

3. The method according to claim 2, characterized in that The qualitative comparative analysis based on the historical partial credit investigation text data of the multiple objects to obtain the target risk evaluation index and the evaluation value corresponding to the target risk evaluation index includes: In response to the user's index selection operation on the historical partial credit investigation text data of the multiple objects, obtaining a first risk assessment index; In response to a user's voting operation on the first risk assessment indicator, determining a second risk assessment indicator based on a voting result; Performing a hierarchical analysis on the second risk assessment indicator to obtain a target risk assessment indicator and a weight corresponding to the target risk assessment indicator; An evaluation value corresponding to the target risk evaluation indicator is determined based on the weight corresponding to the target risk evaluation indicator.

4. The method according to claim 3, characterized in that The step of performing a hierarchical analysis on the second risk assessment indicator to obtain a target risk assessment indicator and a weight corresponding to the target risk assessment indicator includes: The second risk assessment index is hierarchically divided to obtain risk assessment indexes in a target layer, risk assessment indexes in a criterion layer, and risk assessment indexes in a solution layer; Determine a judgment matrix based on the risk assessment indicators in the target layer, the risk assessment indicators in the criterion layer, and the risk assessment indicators in the solution layer; A target risk assessment indicator and a weight corresponding to the target risk assessment indicator are determined based on the judgment matrix.

5. The method according to claim 1, characterized in that The steps of constructing the integrated learning scorecard model include: Acquiring historical full credit text data of multiple objects, wherein the data missing rate of the historical full credit text data is less than a preset missing rate threshold; Performing chi-square binning based on the historical full amount of credit report text data of the multiple objects to obtain binning results; Performing evidence weight conversion on the binning results to obtain the evidence weight value corresponding to the binning results; Based on the evidence weight values ​​corresponding to the binning results, the initial machine learning model is trained to obtain an integrated learning scorecard model.

6. The method according to claim 5, characterized in that The initial machine learning model is trained based on the evidence weight value corresponding to the binning result to obtain an integrated learning scorecard model, including: Based on the evidence weight values ​​corresponding to the binning results, the gradient boosting decision tree is trained to obtain an integrated learning scorecard model.

7. The method according to claim 6, characterized in that The step of training the gradient boosting decision tree based on the weight of evidence corresponding to the binning result to obtain an integrated learning scorecard model includes: Inputting the evidence weight value corresponding to the binning result into the gradient boosting decision tree to obtain decision tree information, wherein the decision tree information includes the depth of the tree, the number of leaf nodes and the output value of the leaf nodes; The decision tree information is input into a loss function with a regularization term to obtain a model loss value, and the model parameters of the gradient boosting decision tree are adjusted based on the model loss value until the training stop condition is met to obtain an integrated learning scorecard model.

8. The method according to claim 5, characterized in that After performing ensemble learning training on the initial machine learning model based on the evidence weight value corresponding to the binning result to obtain the ensemble learning scorecard model, the method further includes: Performing parameter tuning on the integrated learning scorecard model to obtain an integrated learning scorecard model after parameter tuning; Accordingly, the credit report text data of the target object is input into the pre-built integrated learning scorecard model to obtain the integrated learning risk prediction data of the target object, including: The credit report text data of the target object is input into the integrated learning scorecard model after parameter tuning to obtain the integrated learning risk prediction data of the target object.

9. The method according to claim 1, characterized in that: The determining the target risk prediction data of the target object based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object includes: Determine the weight information of the expert experience risk prediction data and the weight information corresponding to the ensemble learning risk prediction data; Determine first risk prediction data based on the expert experience risk prediction data of the target object and weight information of the expert experience risk prediction data; Determine second risk prediction data based on the ensemble learning risk prediction data of the target object and weight information corresponding to the ensemble learning risk prediction data; The target risk prediction data of the target object is determined based on the first risk prediction data and the second risk prediction data.

10. The method according to claim 9, characterized in that The determining of the weight information of the expert experience risk prediction data and the weight information corresponding to the ensemble learning risk prediction data includes: Obtain the data missing rate of historical incomplete credit report text data; Determining the weight information of the expert experience risk prediction data based on the data missing rate of the historical incomplete credit investigation text data; Obtain the data missing rate of the entire historical credit report text data; The weight information corresponding to the integrated learning risk prediction data is determined based on the data missing rate of the historical full credit text data.

11. The method according to claim 9, characterized in that The determining of the weight information of the expert experience risk prediction data and the weight information corresponding to the ensemble learning risk prediction data includes: Obtain the prediction accuracy of the expert experience scorecard model; Determining weight information of the expert experience risk prediction data based on the prediction accuracy of the expert experience scorecard model; Get the prediction accuracy of the ensemble learning scorecard model; The weight information corresponding to the ensemble learning risk prediction data is determined based on the prediction accuracy of the ensemble learning scorecard model.

12. A credit investigation data processing device, characterized in that: include: A credit information data acquisition module is used to acquire the credit information text data of the target object; An expert experience scoring module is used to input the credit report text data of the target object into a pre-built expert experience scoring card model to obtain the expert experience risk prediction data of the target object; An integrated learning scoring module, used to input the credit report text data of the target object into a pre-built integrated learning scoring card model to obtain the integrated learning risk prediction data of the target object; The target risk prediction data determination module is used to determine the target risk prediction data of the target object based on the expert experience risk prediction data of the target object and the integrated learning risk prediction data of the target object.

13. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the credit data processing method according to any one of claims 1 to 11 when executing the computer program.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the credit data processing method as described in any one of claims 1-11 is implemented.

15. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the computer program implements the credit data processing method according to any one of claims 1 to 11.

Images