CN111476658A - Loan continuous overdue prediction method and device - Google Patents
Loan continuous overdue prediction method and device Download PDFInfo
- Publication number
- CN111476658A CN111476658A CN202010283945.4A CN202010283945A CN111476658A CN 111476658 A CN111476658 A CN 111476658A CN 202010283945 A CN202010283945 A CN 202010283945A CN 111476658 A CN111476658 A CN 111476658A
- Authority
- CN
- China
- Prior art keywords
- loan
- overdue
- continuous
- prediction model
- information
- 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; CALCULATING OR 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; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
Abstract
The invention provides a method and a device for predicting continuous overdue of a loan, wherein the method comprises the following steps: obtaining loan overdue evaluation information of a client; preprocessing the loan overdue evaluation information to obtain loan characteristic data; inputting the loan characteristic data into a loan continuous overdue prediction model, and outputting the prediction result of the loan continuous overdue of the client; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label. The device is used for executing the method. The method and the device for predicting the continuous overdue of the loan, provided by the embodiment of the invention, can improve the accuracy of the continuous overdue prediction of the loan.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method and a device for predicting continuous overdue of a loan.
Background
With the gradual change of the consumption habits of users, the loan amount is continuously increased, and the overdue risk of the loan is also continuously increased. Therefore, the party offering the loan needs to manage the risk of the loan being overdue.
Currently, post-loan management in the industry relies on expert experience to make decisions, and subjective factors have a great influence. The monitoring index of the loan is single, mainly the post-event indexes such as overdue days, overdue amount and the like, so that the bad loan is not found in time, and the post-event wind control management of the loan is passive.
Disclosure of Invention
In view of the problems in the prior art, embodiments of the present invention provide a method and an apparatus for predicting a continuous overdue loan, which can at least partially solve the problems in the prior art.
On one hand, the invention provides a method for predicting continuous overdue of a loan, which comprises the following steps:
obtaining loan overdue evaluation information of a client;
preprocessing the loan overdue evaluation information to obtain loan characteristic data;
inputting the loan characteristic data into a loan continuous overdue prediction model, and outputting the prediction result of the loan continuous overdue of the client; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
In another aspect, the present invention provides a device for predicting continuous overdue of a loan, comprising:
the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring loan overdue evaluation information of a client;
the preprocessing unit is used for preprocessing the loan overdue evaluation information to obtain loan characteristic data;
the prediction unit is used for inputting the loan feature data into a loan continuous overdue prediction model and outputting the prediction result of the loan continuous overdue of the client; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
In another aspect, the present invention provides an electronic device, which includes a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method for predicting continuous overdue loan in any of the embodiments.
In yet another aspect, the present invention provides a computer readable storage medium, on which a computer program is stored, the computer program, when executed by a processor, implementing the steps of the method for predicting continuous overdue of a loan according to any of the embodiments described above.
The method and the device for predicting the continuous overdue loan, provided by the embodiment of the invention, have the advantages that the loan overdue evaluation information of a client is obtained, the loan characteristic data is obtained by preprocessing the loan overdue evaluation information, the loan characteristic data is input into a loan continuous overdue prediction model, the prediction result of the continuous overdue loan of the client is output, and the accuracy of the loan continuous overdue prediction can be improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts. In the drawings:
fig. 1 is a flowchart illustrating a method for predicting continuous overdue of a loan according to an embodiment of the invention.
Fig. 2 is a flowchart illustrating a method for predicting continuous overdue of a loan according to another embodiment of the invention.
Fig. 3 is a flowchart illustrating a method for predicting continuous overdue of a loan according to another embodiment of the invention.
Fig. 4 is a schematic structural diagram of a device for predicting continuous overdue loan, according to an embodiment of the invention.
Fig. 5 is a schematic structural diagram of a device for predicting continuous overdue loan, according to another embodiment of the invention.
Fig. 6 is a schematic structural diagram of a device for predicting continuous overdue loan, according to another embodiment of the invention.
Fig. 7 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention. It should be noted that the embodiments and features of the embodiments in the present application may be arbitrarily combined with each other without conflict.
Fig. 1 is a schematic flow chart of a method for predicting continuous overdue of a loan according to an embodiment of the present invention, and as shown in fig. 1, the method for predicting continuous overdue of a loan according to an embodiment of the present invention includes:
s101, obtaining loan overdue evaluation information of a customer;
specifically, the loan overdue evaluation information of the customer is information for reflecting whether the customer's loan will be continuously overdue. The loan overdue assessment information of the client can comprise information which can be directly obtained by age, gender, education degree, industry, loan products and the like according to different sources, and can also comprise information which is obtained by data processing such as the proportion of the loan account age to the loan term, the amount of money inflow in the current term, the difference value between the amount of money inflow in the current term and the amount of repayment and the like. The server may obtain loan overdue evaluation information for the customer. The amount of information included in the loan overdue evaluation information of the customer is set according to actual needs, and the embodiment of the invention is not limited. It can be understood that when data is missing in the loan overdue evaluation information of the customer, the data can be supplemented by a random forest algorithm, an average value method or data of similar customers, and the data is selected according to actual needs, which is not limited in the embodiment of the invention. The implementation subject of the prediction method of continuous overdue loan provided by the embodiment of the invention includes but is not limited to a server.
For example, table 1 is a customer marriage status information table, and as shown in table 1, the marriage status of the customer E is unknown, and it is statistically known that the number of customers whose marriage status is married is larger than the number of customers who are not married, so that the customer E whose marriage status is missing is complemented with a large number of category values using the column data, and the marriage status of the complemented customer E is married.
TABLE 1 customer marital status information Table
| Customer | Marital status |
| A | Wedding |
| B | Wedding |
| C | Wedding |
| D | Unmarried |
| E | Is unknown |
S102, preprocessing the loan overdue evaluation information to obtain loan feature data;
specifically, after the loan overdue evaluation information is obtained, the server preprocesses the loan overdue evaluation information, converts the loan overdue evaluation information into numerical data, and obtains loan feature data. Wherein, the numerical value can be directly reserved for the information such as the total amount of financial products, the balance of various financial products and the like; information such as gender, industry, education level and the like can be converted into numerical data by one-hot coding and the like. The specific process of converting the loan overdue evaluation information into numerical data is set according to actual needs, and the embodiment of the invention is not limited.
For example, for gender, a value corresponding to males may be set to 1 and a value corresponding to females may be set to 0.
For example, the education level of the customer can be converted into numerical data by one-hot coding. The education level of the client a was high school, the education level of the client B was university department, and the numerical data into which the education levels of the client a and the client B were converted are shown in table 2.
TABLE 2 numerical data of education level of client
| Customer | Go to school | Primary school | Middle school | High school | University discipline | University's this branch of academic or vocational study | Research student |
| A | 0 | 0 | 0 | 1 | 0 | 0 | 0 |
| B | 0 | 0 | 0 | 0 | 0 | 1 | 0 |
For example, the loan account age and the loan term for a customer may be converted into a loan account age to loan term ratio. As shown in table 3, if the loan account age of client a is 12 th and the loan term is 360 th, the proportion of the loan account age of client a to the loan term is 12/360-0.0333. If the credit account of client B is 16 th and the credit term is 24 th, the ratio of the credit account of client B to the credit term is 16/24-0.6667.
TABLE 3 loan account age to loan term ratio
| Customer | Loan account age | Loan term | Loan account age in loan deadline proportion |
| A | 12 | 360 | 0.0333 |
| B | 16 | 24 | 0.6667 |
S103, inputting the loan feature data into a loan continuous overdue prediction model, and outputting a prediction result of the loan continuous overdue of the customer; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
Specifically, after the server obtains the loan feature data, the loan feature data is input into a loan continuous overdue prediction model, and the prediction result of the continuous overdue of the loan of the client can be output through the processing of the loan continuous overdue prediction model, wherein the prediction result of the continuous overdue of the loan is continuous overdue or not, and the fact that the continuous overdue of the loan indicates that the client has a high loan overdue risk requires to strengthen risk management and control on the client. The loan overdue prediction result shows that the client loan overdue risk is low because continuous overdue does not occur. The loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label. And each evaluation sample data in the loan overdue evaluation sample data corresponds to a continuous overdue label of the loan. The loan continuous overdue prediction model is used for predicting whether a client of a certain loan has the continuous overdue condition of the loan.
For example, the loan continuous overdue tag is obtained based on historical loan data for a certain loan of 3 continuous dates. As shown in Table 4, three consecutive months of loan data for loan X, numbered 1-8, were obtained. And setting a continuous overdue label of the loan to be continuous two overdues, and abandoning the evaluation sample data corresponding to the number 1 for the loan data with the number 1 because the loan data with the number 1 has been continuous 3 overdues, namely the loan data with the number 1 has been continuous 2 overdues by the set current month. For the loan data with the number of 2, since 2 continuous overdue dates have been reached by the set current month, the evaluation sample data corresponding to the number of 2 is abandoned. For the loan data with the number of 5, since 2 continuous overdue dates from the current month are set, the continuous overdue tag of the loan corresponding to the evaluation sample data with the number of 5 is set to be 1, which indicates that the loan is continuously overdue for 2 times. For the loan data with the rest numbers, as the condition that 2 continuous overdue exists, the continuous overdue label of the loan corresponding to the evaluation sample data corresponding to the rest numbers is set to be 0, which indicates that 2 continuous overdue times do not exist. And obtaining evaluation sample data corresponding to the continuous overdue label of the loan according to the relevant information of the client corresponding to the loan term number corresponding to the continuous overdue label of the loan.
TABLE 4 continuous overdue tag for loan
| Numbering | Whether the previous month is overdue | Whether the current month is overdue | Whether the next month is overdue | Label (R) |
| 1 | 1 | 1 | 1 | Discard the |
| 2 | 1 | 1 | 0 | Discard the |
| 3 | 1 | 0 | 1 | 0 |
| 4 | 1 | 0 | 0 | 0 |
| 5 | 0 | 1 | 1 | 1 |
| 6 | 0 | 1 | 0 | 0 |
| 7 | 0 | 0 | 0 | 0 |
| 8 | 0 | 0 | 1 | 0 |
The method for predicting the continuous overdue of the loan, provided by the embodiment of the invention, has the advantages that the loan overdue evaluation information of a client is obtained, the loan overdue evaluation information is preprocessed, the loan characteristic data is obtained, the loan characteristic data is input into a loan continuous overdue prediction model, the prediction result of the continuous overdue of the loan of the client is output, and the accuracy of prediction of the continuous overdue of the loan can be improved. In addition, the continuous overdue of the loan is predicted, the condition that the loan is overdue once occasionally is filtered, and the loan with higher risk can be predicted more accurately.
Fig. 2 is a flowchart illustrating a method for predicting the continuous overdue of a loan according to another embodiment of the present invention, and as shown in fig. 2, the step of training the model for predicting the continuous overdue of the loan based on the sample data for evaluating the loan and the predetermined label for the loan continuous overdue includes:
s201, preprocessing the sample data of the loan overdue evaluation to obtain a sample feature set, and dividing the sample feature set into a training set, a verification set and a test set;
specifically, loan data of N consecutive months of a certain loan may be collected and obtained, the loan data of the N consecutive months may be labeled to obtain a first number of labels, a first number of evaluation sample data may be obtained based on client-related information of a loan term corresponding to each label, each evaluation sample data corresponds to one loan continuous overdue label, and the first number of evaluation sample data constitutes the loan overdue evaluation sample data. The server preprocesses the loan overdue evaluation sample data to obtain a sample feature set, wherein the sample feature set comprises a first number of sample feature data. The server divides the sample feature set into a training set, a validation set, and a test set, for example, attributing 70% of the first number of sample feature data to the training set, attributing 20% of the first number of sample feature data to the validation set, and attributing the remaining sample feature data to the test set. The specific value of N is set according to actual needs, and the embodiment of the present invention is not limited. The preprocessing process of the loan overdue evaluation sample data is similar to the preprocessing process of the loan overdue evaluation information, and is not repeated here. It is understood that each evaluation sample data includes the same category of information as the category of information included in the loan overdue evaluation information of the customer.
For example, the method may include collecting loan data obtained by obtaining 8 consecutive loan dates of the loan Y before the predicted time point, applying a label of 2 consecutive overdue loans for each of three consecutive loan dates, obtaining a label of 2 consecutive overdue loans corresponding to each of 6 consecutive loans from the preset time point, obtaining one evaluation sample data by using the related information of the number of loan dates corresponding to each label of 2 consecutive overdue loans, and assuming that a second number of evaluation sample data constituting the loan overdue evaluation sample data of the loan Y is obtained in total.
S202, training to obtain a to-be-determined loan continuous overdue prediction model according to the training set, the loan continuous overdue labels corresponding to the training set and a gradient elevator algorithm model;
specifically, the server inputs the sample characteristic data concentrated in training and the continuous overdue loan labels corresponding to the sample characteristic data into a gradient elevator algorithm model to obtain a prediction result of the sample characteristic data, then corrects the gradient elevator algorithm model through a residual error between the prediction result of the sample characteristic data and the continuous loan overdue loan labels corresponding to the sample characteristic data, and continuously and circularly iterates until the iteration times reach preset times to obtain the continuous loan overdue prediction model to be determined. The preset times are set according to actual needs, and the embodiment of the invention is not limited.
The Gradient Boosting Machine (GBM) algorithm is an integrated algorithm, and a plurality of weak learners are integrated together to form a strong learner. During the training process of the gradient elevator algorithm model, firstly, default is usedTo construct an initial tree model F1(x) For training data (x)i,yi)，xiAs sample feature data, yiFor the loan continuous overdue label corresponding to the sample characteristic data, x is addediInput to the initial Tree model F1(x) Output predicted value is y'iY'i＝F1(xi)。yiAnd y'iWith a residual y betweeni-y′iI.e. L (y)i,F1(xi))＝yi-y′i，L(y,y′i) To find the optimal tree model F (x) from the training set, the loss function L (y, F (x)) is minimized2(x)＝F1(x)+h(x)，F2(x) For the new tree model, h (x) is the added estimator. The best model effect is achieved when h (x) should satisfy h (x) y-F (x), so h (x) and the residual y-F1(x) Fitting, namely continuously fitting the residual error of the previous tree model, correcting the tree model obtained in the previous time, and obtaining a final tree model F (x) after M iterations. Wherein the value of the negative gradient of the loss function at the current tree model is used as an approximation of the residual error. To prevent under-fitting and over-fitting, the number of iterations M is generally selected to be a moderate value, such as 100, and is determined empirically, and the embodiment of the present invention is not limited thereto.
S203, verifying the to-be-determined loan continuous overdue prediction model according to the verification set and the loan continuous overdue label corresponding to the verification set;
specifically, after the server obtains the model for continuous overdue prediction of the to-be-determined loan, the effect of the model for continuous overdue prediction of the to-be-determined loan may be verified through the verification set, the sample feature data of the verification set and the label for continuous overdue of the loan corresponding to the sample feature data are input to the model for continuous overdue prediction of the to-be-determined loan, the prediction result of each sample feature data of the verification set is obtained, and the model for continuous overdue prediction of the to-be-determined loan is verified based on the prediction result of each sample feature data of the verification set and the label for continuous overdue of the loan corresponding to each sample feature data. If the to-be-determined loan continuous overdue prediction model passes the verification, testing the to-be-determined loan continuous overdue prediction model by using the test set; and if the model for predicting the continuous overdue loan to be determined does not pass the verification, adjusting all parameters of the gradient elevator algorithm model, and performing model training again.
S204, if the to-be-determined loan continuous overdue prediction model is judged to pass the verification, testing the to-be-determined loan continuous overdue prediction model according to the test set and the loan continuous overdue label corresponding to the test set;
specifically, the server judges whether the to-be-determined loan continuous overdue prediction model passes verification, if the to-be-determined loan continuous overdue prediction model passes verification, the characteristic data of each sample in the test set is input into the to-be-determined loan continuous overdue prediction model to obtain the prediction result of the characteristic data of each sample in the test set, the prediction result of the characteristic data of each sample in the test set is compared with the loan continuous overdue label corresponding to the characteristic data of each sample, the prediction accuracy of the to-be-determined loan continuous overdue prediction model can be obtained, if the prediction accuracy of the to-be-determined loan continuous overdue prediction model is greater than or equal to the accuracy threshold, the to-be-determined loan continuous overdue prediction model passes test, and if the prediction accuracy of the to-be-determined loan continuous overdue prediction model is smaller than the accuracy threshold, the pending loan continuous overdue prediction model fails the test. The accuracy threshold is set according to actual experience, and the embodiment of the present invention is not limited.
For example, the test set has 20 sample feature data, the 20 sample feature data are input into the to-be-determined loan continuous overdue prediction model to obtain the prediction results of the 20 sample feature data, the prediction results of the 20 sample feature data are respectively compared with loan continuous overdue labels corresponding to the 20 sample feature data, and as a result, the prediction results of 18 sample feature data are the same as the corresponding loan continuous overdue labels, so that the accuracy of prediction of the to-be-determined loan continuous overdue prediction model is 18/20-90%.
And S205, if the fact that the to-be-determined loan continuous overdue prediction model passes the test is judged, taking the to-be-determined loan continuous overdue prediction model as the loan continuous overdue prediction model.
Specifically, after testing the to-be-determined loan continuous overdue prediction model, the server judges whether the to-be-determined loan continuous overdue prediction model passes the test, and if the to-be-determined loan continuous overdue prediction model passes the test, the to-be-determined loan continuous overdue prediction model is used as the loan continuous overdue prediction model.
Fig. 3 is a flowchart of a method for predicting continuous overdue of a loan according to another embodiment of the present invention, and as shown in fig. 3, on the basis of the foregoing embodiments, the verifying the model for predicting continuous overdue of a loan to be determined according to the verification set and the continuous overdue tag of the loan corresponding to the verification set includes:
s2031, respectively inputting the sample characteristic data of the verification set into the continuous overdue prediction model of the loan to be determined, and outputting the prediction result of each sample characteristic data;
specifically, the server inputs each sample feature data of the verification set into the model for predicting the loan continuity to be determined, and may output the prediction result of each sample feature data of the verification set.
S2032, obtaining an ROC curve according to the prediction result of each sample characteristic data of the verification set and the loan continuous overdue label corresponding to each sample characteristic data;
specifically, the server compares the prediction result of each sample characteristic data of the verification set with the loan continuous overdue tag corresponding to each sample characteristic data, and may draw an roc (receiveoperating characteristic) curve according to the comparison result.
TABLE 5 confusion number table
For example, the continuous overdue tag of the loan is represented by 1, and 0 represents no continuous overdue. In the prediction results, 1 indicates continuous overdue, and 0 indicates no continuous overdue. Assuming that the verification set comprises 40 sample feature data, comparing the prediction results of the 40 sample feature data with the corresponding loan continuous overdue tags, and determining the true yang number, the false yin number and the true yin number as shown in table 5, where the true yang number is 5, the false yang number is 2, the false yin number is 2 and the true yin number is 31, then calculating a true yang rate of 5/(5+2) of 0.714, a false yang rate of 2/(2+31) of 0.061, and the server may draw an ROC curve according to the true yang rate and the false yang rate.
S2033, verifying the to-be-determined loan continuous overdue prediction model according to the AUC value corresponding to the ROC curve.
Specifically, after the ROC curve is drawn by the server, an AUC (area Under customer) value can be obtained by calculating the area enclosed by the ROC curve and a coordinate axis, if the absolute value of the difference between the AUC value and 1 is smaller than a preset value, the to-be-determined loan continuous overdue prediction model passes verification, and if the absolute value of the difference between the AUC value and 1 is larger than or equal to the preset value, the to-be-determined loan continuous overdue prediction model does not pass verification.
On the basis of the above embodiments, further, the loan overdue evaluation information includes six types of information, i.e., basic customer information, customer property information, customer loan information, transaction information, risk information, and similar customer information.
Specifically, the loan overdue evaluation information of the customer may be classified into six types: the system comprises client basic information, client property information, client loan information, transaction information, risk information and the like client information. The basic information of the client comprises information such as age, gender, education level, marital status, industry and the like; the client asset information comprises information such as the total amount of financial products held by the current client, the balance of financial products of various categories and the like; the client loan information comprises information of balance of various loan products, loan account age in loan deadline proportion and the like; the transaction information comprises information such as the number of fund inflows/outflows, the total amount of fund inflows/outflows and the like of the current month of the client repayment account; the risk information comprises information such as the proportion of the amount of the credit not yet cleared to the total amount of the credit of the client, the ratio of the current credit balance to the financial assets, the current loan recombination number and the like; the similar customer information comprises information of batch application, batch overdue, same fund source account and the like, and is used for considering the loan characteristics of whether the customers have batch false loans, and the loan customers under the same developer may have the false loans.
The information of whether the loan is overdue or not is provided by the six types of information from different dimensions, the comprehensiveness of the overdue characteristic of the loan is improved, the training of the continuous overdue loan prediction model is performed by using the six types of information, the accuracy of the continuous loan overdue prediction model can be improved, and the accuracy of continuous loan overdue prediction is further improved.
The following describes a specific embodiment of the implementation of the method for predicting the continuous overdue loan according to the embodiment of the present invention.
A commercial bank a now predicts that a loan is two consecutive overdues for a customer of a property loan. Firstly, a model for predicting two times of continuous overdue of a certain property loan is required to be established, loan data of all clients of a certain property loan for eight consecutive months are collected, tagging is carried out by referring to a table 4, data which cannot be tagged is discarded, if Q tags are obtained, then client basic information, client property information, client loan information, transaction information, risk information and similar client information of clients with the number of loan periods corresponding to each tag of two times of continuous overdue are collected, and Q evaluation sample data are obtained in total and serve as loan overdue evaluation sample data of the certain property loan. The method comprises the steps that the server preprocesses loan overdue evaluation sample data of a certain property loan to obtain a sample feature set of the loan of the certain property, the sample feature set of the loan of the certain property comprises Q sample feature data, the Q sample feature data are divided into a training set, a verification set and a test set, the training set comprises 0.7Q sample feature data, the verification set comprises 0.2Q sample feature data, and the test set comprises 0.1Q sample feature data. And training to obtain a model for predicting the continuous overdue loan of a certain house loan to be determined according to the 0.7Q sample characteristic data, the label of the loan for two continuous overdue times corresponding to the 0.7Q sample characteristic data and the gradient elevator algorithm model. And verifying the continuous overdue prediction model of the loan to be determined of a certain property loan by using the 0.2Q sample characteristic data and the label of the loan corresponding to the 0.2Q sample characteristic data and two continuous overdue times. After the to-be-determined loan continuous overdue prediction model of the property loan passes verification, the to-be-determined loan continuous overdue prediction model of the property loan is tested by using the 0.1Q sample characteristic data and the label of the property loan corresponding to the 0.1Q sample characteristic data and used for two times of continuous overdue. And after the model for predicting the continuous overdue loan of the house loan to be determined passes the verification, taking the model for predicting the continuous overdue loan of the house loan to be determined as a model for predicting the continuous twice overdue loan of the house.
When the client C is predicted to be overdue twice continuously, the client C is collected with the client basic information, the client property information, the client loan information, the transaction information, the risk information and the similar client information. The basic client information, the property information, the loan information, the transaction information, the risk information and the similar client information of the client C are preprocessed to obtain the loan characteristic data of the client C, and the loan characteristic data of the client C is input into a continuous overdue twice-prediction model of a property loan, so that the overdue prediction result of the client C loan can be output. If the output result shows that the client C can continuously overdue the property loan twice, the risk control of the client C needs to be strengthened, and the client C is focused when the risk control is carried out on the property loan.
Fig. 4 is a schematic structural diagram of a prediction apparatus of continuous overdue loan according to an embodiment of the present invention, as shown in fig. 4, on the basis of the foregoing embodiments, further, the prediction apparatus of continuous overdue loan according to an embodiment of the present invention includes an obtaining unit 401, a preprocessing unit 402, and a prediction unit 403, where:
the obtaining unit 401 is configured to obtain loan overdue evaluation information of a customer; the preprocessing unit 402 is configured to preprocess the loan overdue evaluation information to obtain loan feature data; the prediction unit 403 is configured to input the loan feature data into a loan continuous overdue prediction model, and output a prediction result of the loan continuous overdue of the customer; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
Specifically, the loan overdue evaluation information of the customer is information for reflecting whether the customer's loan will be continuously overdue. The loan overdue assessment information of the client can comprise information which can be directly obtained by age, gender, education degree, industry, loan products and the like according to different sources, and can also comprise information which is obtained by data processing such as the proportion of the loan account age to the loan term, the amount of money inflow in the current term, the difference value between the amount of money inflow in the current term and the amount of repayment and the like. The acquisition unit 401 may acquire loan overdue evaluation information of the customer. The amount of information included in the loan overdue evaluation information of the customer is set according to actual needs, and the embodiment of the invention is not limited. It can be understood that when data is missing in the loan overdue evaluation information of the customer, the data can be supplemented by a random forest algorithm, an average value method or data of similar customers, and the data is selected according to actual needs, which is not limited in the embodiment of the invention.
After obtaining the loan overdue evaluation information, the preprocessing unit 402 may preprocess the loan overdue evaluation information, convert the loan overdue evaluation information into numerical data, and obtain loan feature data. Wherein, the numerical value can be directly reserved for the information such as the total amount of financial products, the balance of various financial products and the like; information such as gender, industry, education level and the like can be converted into numerical data by one-hot coding and the like. The specific process of converting the loan overdue evaluation information into numerical data is set according to actual needs, and the embodiment of the invention is not limited.
After the loan feature data is obtained, the prediction unit 403 inputs the loan feature data into a loan continuous overdue prediction model, and the processing of the loan continuous overdue prediction model may output a prediction result of the continuous overdue of the loan of the customer, where the prediction result of the continuous overdue of the loan is continuous overdue or may not be continuous overdue, and the continuous overdue of the loan is continuous overdue, which indicates that the customer has a high loan overdue risk and needs to strengthen the management and control risk of the customer. The loan overdue prediction result shows that the client loan overdue risk is low because continuous overdue does not occur. The loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label. And each evaluation sample data in the loan overdue evaluation sample data corresponds to a continuous overdue label of the loan. The loan continuous overdue prediction model is used for predicting whether a client of a certain loan has the continuous overdue condition of the loan.
The prediction device for the continuous overdue loan, provided by the embodiment of the invention, is used for acquiring the loan overdue evaluation information of a client, preprocessing the loan overdue evaluation information to acquire loan characteristic data, inputting the loan characteristic data into the loan continuous overdue prediction model, outputting the prediction result of the continuous overdue loan of the client, and improving the accuracy of prediction of the continuous overdue loan. In addition, the continuous overdue of the loan is predicted, the condition that the loan is overdue once occasionally is filtered, and the loan with higher risk can be predicted more accurately.
Fig. 5 is a schematic structural diagram of a prediction apparatus of continuous overdue loan according to another embodiment of the present invention, as shown in fig. 5, based on the foregoing embodiments, further including a dividing unit 404, a training unit 405, a verification unit 406, a testing unit 407, and a determining unit 408, where:
the dividing unit 404 is configured to preprocess the loan overdue evaluation sample data to obtain a sample feature set, and divide the sample feature set into a training set, a verification set, and a test set; the training unit 405 is configured to train to obtain a to-be-determined loan continuous overdue prediction model according to the training set, the loan continuous overdue label corresponding to the training set, and the gradient elevator algorithm model; the verification unit 406 is configured to verify the to-be-determined loan continuous overdue prediction model according to the verification set and the loan continuous overdue tag corresponding to the verification set; the testing unit 407 is configured to test the to-be-determined loan continuous overdue prediction model according to the test set and the loan continuous overdue tag corresponding to the test set after determining that the to-be-determined loan continuous overdue prediction model passes verification; the judging unit 408 is configured to, after judging that the to-be-determined loan continuous overdue prediction model passes the test, take the to-be-determined loan continuous overdue prediction model as the loan continuous overdue prediction model.
Specifically, loan data of N consecutive months of a certain loan may be collected and obtained, the loan data of the N consecutive months may be labeled to obtain a first number of labels, a first number of evaluation sample data may be obtained based on client-related information of a loan term corresponding to each label, each evaluation sample data corresponds to one loan continuous overdue label, and the first number of evaluation sample data constitutes the loan overdue evaluation sample data. The partitioning unit 404 preprocesses the loan overdue evaluation sample data to obtain a sample feature set, where the sample feature set includes a first number of sample feature data. The server divides the sample feature set into a training set, a validation set, and a test set, for example, attributing 70% of the first number of sample feature data to the training set, attributing 20% of the first number of sample feature data to the validation set, and attributing the remaining sample feature data to the test set. The specific value of N is set according to actual needs, and the embodiment of the present invention is not limited. The preprocessing process of the loan overdue evaluation sample data is similar to the preprocessing process of the loan overdue evaluation information, and is not repeated here. It is understood that each evaluation sample data includes the same category of information as the category of information included in the loan overdue evaluation information of the customer.
The training unit 405 inputs the sample characteristic data concentrated in training and the continuous loan overdue labels corresponding to the sample characteristic data into the gradient elevator algorithm model to obtain a prediction result of the sample characteristic data, then corrects the gradient elevator algorithm model through a residual error between the prediction result of the sample characteristic data and the continuous loan overdue labels corresponding to the sample characteristic data, and continuously and circularly iterates until the iteration number reaches a preset number, so that the continuous loan overdue prediction model to be determined is obtained. The preset times are set according to actual needs, and the embodiment of the invention is not limited.
After obtaining the model for continuous overdue prediction of the loan to be determined, the verifying unit 406 may verify the effect of the model for continuous overdue prediction of the loan to be determined by using the verification set, input the sample feature data of the verification set and the label for continuous overdue of the loan corresponding to the sample feature data to the model for continuous overdue prediction of the loan to be determined, obtain the prediction result of each sample feature data of the verification set, and verify the model for continuous overdue prediction of the loan to be determined based on the prediction result of each sample feature data of the verification set and the label for continuous overdue of the loan corresponding to each sample feature data. If the to-be-determined loan continuous overdue prediction model passes the verification, testing the to-be-determined loan continuous overdue prediction model by using the test set; and if the model for predicting the continuous overdue loan to be determined does not pass the verification, adjusting all parameters of the gradient elevator algorithm model, and performing model training again.
The testing unit 407 may determine whether the to-be-determined loan continuous overdue prediction model passes the verification, if the to-be-determined loan continuous overdue prediction model passes the verification, input each sample feature data in the testing set to the to-be-determined loan continuous overdue prediction model, obtain a prediction result of each sample feature data in the testing set, compare the prediction result of each sample feature data in the testing set with a loan continuous overdue tag corresponding to each sample feature data, may obtain the accuracy of the to-be-determined loan continuous overdue prediction model prediction, if the accuracy of the to-be-determined loan continuous overdue prediction model prediction is greater than or equal to an accuracy threshold, the to-be-determined loan continuous overdue prediction model passes the testing, if the accuracy of the to-be-determined loan continuous overdue prediction model prediction is less than the accuracy threshold, the pending loan continuous overdue prediction model fails the test. The accuracy threshold is set according to actual experience, and the embodiment of the present invention is not limited.
After the to-be-determined loan continuous overdue prediction model is tested, the determining unit 408 determines whether the to-be-determined loan continuous overdue prediction model passes the test, and if the to-be-determined loan continuous overdue prediction model passes the test, the to-be-determined loan continuous overdue prediction model is used as the loan continuous overdue prediction model.
Fig. 6 is a schematic structural diagram of a device for predicting continuous overdue of a loan according to another embodiment of the present invention, as shown in fig. 6, and based on the foregoing embodiments, the verifying unit 406 further includes an outputting sub-unit 4061, an obtaining sub-unit 4062, and a determining sub-unit 4063, where:
the output sub-unit 4061 is configured to input each sample feature data of the verification set to the to-be-determined loan continuous overdue prediction model, and output a prediction result of each sample feature data; the obtaining subunit 4062 is configured to obtain an ROC curve according to the prediction result of each sample feature data of the verification set and the loan continuous overdue tag corresponding to each sample feature data; the determining subunit 4063 is configured to verify the to-be-determined loan continuous overdue prediction model according to an AUC value corresponding to the ROC curve.
Specifically, the output sub-unit 4061 inputs each sample feature data of the verification set into the model for predicting the loan continuity to be determined, and may output the prediction result of each sample feature data of the verification set.
The obtaining subunit 4062 compares the prediction result of each sample feature data of the verification set and the loan continuous overdue label corresponding to each sample feature data, and may draw an ROC curve according to the comparison result.
After the ROC curve is drawn, the determining subunit 4063 calculates an AUC value obtained by calculating an area enclosed by the ROC curve and a coordinate axis, if an absolute value of a difference between the AUC value and 1 is smaller than a preset value, the to-be-determined loan continuous overdue prediction model passes verification, and if the absolute value of a difference between the AUC value and 1 is greater than or equal to the preset value, the to-be-determined loan continuous overdue prediction model does not pass verification.
On the basis of the above embodiments, further, the loan overdue evaluation information includes six types of information, i.e., basic customer information, customer property information, customer loan information, transaction information, risk information, and similar customer information.
Specifically, the loan overdue evaluation information of the customer may be classified into six types: the system comprises client basic information, client property information, client loan information, transaction information, risk information and the like client information. The basic information of the client comprises information such as age, gender, education level, marital status, industry and the like; the client asset information comprises information such as the total amount of financial products held by the current client, the balance of financial products of various categories and the like; the client loan information comprises information of balance of various loan products, loan account age in loan deadline proportion and the like; the transaction information comprises information such as the number of fund inflows/outflows, the total amount of fund inflows/outflows and the like of the current month of the client repayment account; the risk information comprises information such as the proportion of the amount of the credit not yet cleared to the total amount of the credit of the client, the ratio of the current credit balance to the financial assets, the current loan recombination number and the like; the similar customer information comprises information of batch application, batch overdue, same fund source account and the like, and is used for considering the loan characteristics of whether the customers have batch false loans, and the loan customers under the same developer may have the false loans.
The embodiment of the apparatus provided in the embodiment of the present invention may be specifically configured to execute the processing flows of the above method embodiments, and the functions of the apparatus are not described herein again, and refer to the detailed description of the above method embodiments.
Fig. 7 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 7, the electronic device may include: a processor (processor)701, a communication Interface (Communications Interface)702, a memory (memory)703 and a communication bus 704, wherein the processor 701, the communication Interface 702 and the memory 703 complete communication with each other through the communication bus 704. The processor 701 may call logic instructions in the memory 703 to perform the following method: obtaining loan overdue evaluation information of a client; preprocessing the loan overdue evaluation information to obtain loan characteristic data; inputting the loan characteristic data into a loan continuous overdue prediction model, and outputting the prediction result of the loan continuous overdue of the client; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
In addition, the logic instructions in the memory 703 can be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The present embodiment discloses a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the method provided by the above-mentioned method embodiments, for example, comprising: obtaining loan overdue evaluation information of a client; preprocessing the loan overdue evaluation information to obtain loan characteristic data; inputting the loan characteristic data into a loan continuous overdue prediction model, and outputting the prediction result of the loan continuous overdue of the client; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
The present embodiment provides a computer-readable storage medium, which stores a computer program, where the computer program causes the computer to execute the method provided by the above method embodiments, for example, the method includes: obtaining loan overdue evaluation information of a client; preprocessing the loan overdue evaluation information to obtain loan characteristic data; inputting the loan characteristic data into a loan continuous overdue prediction model, and outputting the prediction result of the loan continuous overdue of the client; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In the description herein, reference to the description of the terms "one embodiment," "a particular embodiment," "some embodiments," "for example," "an example," "a particular example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A method for predicting continuous overdue of a loan, comprising:
obtaining loan overdue evaluation information of a client;
preprocessing the loan overdue evaluation information to obtain loan characteristic data;
inputting the loan characteristic data into a loan continuous overdue prediction model, and outputting the prediction result of the loan continuous overdue of the client; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
2. The method of claim 1, wherein the step of training the loan continuous overdue prediction model based on loan overdue evaluation sample data and a predetermined loan continuous overdue label comprises:
preprocessing the sample data of the loan overdue evaluation to obtain a sample feature set, and dividing the sample feature set into a training set, a verification set and a test set;
training to obtain a loan continuous overdue prediction model to be determined according to the training set, the loan continuous overdue label corresponding to the training set and a gradient elevator algorithm model;
verifying the continuous overdue prediction model of the loan to be determined according to the verification set and the continuous overdue label of the loan corresponding to the verification set;
if the to-be-determined loan continuous overdue prediction model is judged to pass the verification, testing the to-be-determined loan continuous overdue prediction model according to the test set and the loan continuous overdue label corresponding to the test set;
and if the to-be-determined loan continuous overdue prediction model passes the test, taking the to-be-determined loan continuous overdue prediction model as the loan continuous overdue prediction model.
3. The method according to claim 2, wherein the verifying the model for predicting loan continuance overdue according to the verification set and the label for loan continuance overdue corresponding to the verification set comprises:
respectively inputting the sample characteristic data of the verification set into the continuous overdue prediction model of the loan to be determined, and outputting the prediction result of each sample characteristic data;
obtaining an ROC curve according to the prediction result of each sample characteristic data of the verification set and the loan continuous overdue label corresponding to each sample characteristic data;
and verifying the continuous overdue prediction model of the loan to be determined according to the AUC value corresponding to the ROC curve.
4. The method according to any one of claims 1 to 3, wherein the loan overdue evaluation information includes six types of information of customer basic information, customer property information, customer loan information, transaction information, risk information, and homogeneous customer information.
5. An apparatus for predicting continuous overdue of a loan, comprising:
the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring loan overdue evaluation information of a client;
the preprocessing unit is used for preprocessing the loan overdue evaluation information to obtain loan characteristic data;
the prediction unit is used for inputting the loan feature data into a loan continuous overdue prediction model and outputting the prediction result of the loan continuous overdue of the client; the loan continuous overdue prediction model is obtained after training based on loan overdue evaluation sample data and a predetermined loan continuous overdue label.
6. The apparatus of claim 5, further comprising:
the system comprises a dividing unit, a data processing unit and a data processing unit, wherein the dividing unit is used for preprocessing the loan overdue evaluation sample data to obtain a sample feature set and dividing the sample feature set into a training set, a verification set and a test set;
the training unit is used for training to obtain a loan continuous overdue prediction model to be determined according to the training set, the loan continuous overdue label corresponding to the training set and a gradient elevator algorithm model;
the verification unit is used for verifying the loan continuous overdue prediction model to be determined according to the verification set and the loan continuous overdue label corresponding to the verification set;
the testing unit is used for testing the continuous overdue prediction model of the loan to be determined according to the test set and the continuous overdue label of the loan corresponding to the test set after judging that the continuous overdue prediction model of the loan to be determined passes the verification;
and the judging unit is used for taking the to-be-determined loan continuous overdue prediction model as the loan continuous overdue prediction model after judging that the to-be-determined loan continuous overdue prediction model passes the test.
7. The apparatus of claim 6, wherein the authentication unit comprises:
the output subunit is used for respectively inputting the sample characteristic data of the verification set to the continuous overdue prediction model of the loan to be determined and outputting the prediction result of each sample characteristic data;
the obtaining subunit is used for obtaining an ROC curve according to the prediction result of each sample characteristic data of the verification set and the loan continuous overdue label corresponding to each sample characteristic data;
and the determining subunit is used for verifying the to-be-determined loan continuous overdue prediction model according to the AUC value corresponding to the ROC curve.
8. The apparatus according to any one of claims 5 to 7, wherein the loan overdue evaluation information includes six types of information of customer basic information, customer property information, customer loan information, transaction information, risk information, and homogeneous customer information.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1 to 4 are implemented when the computer program is executed by the processor.
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 of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010283945.4A CN111476658A (en) | 2020-04-13 | 2020-04-13 | Loan continuous overdue prediction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010283945.4A CN111476658A (en) | 2020-04-13 | 2020-04-13 | Loan continuous overdue prediction method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111476658A true CN111476658A (en) | 2020-07-31 |
Family
ID=71752184
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010283945.4A Pending CN111476658A (en) | 2020-04-13 | 2020-04-13 | Loan continuous overdue prediction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111476658A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112053225A (en) * | 2020-09-02 | 2020-12-08 | 中国银行股份有限公司 | Loan overdue prediction method and device, electronic equipment and storage medium |
| CN112150276A (en) * | 2020-09-30 | 2020-12-29 | 中国银行股份有限公司 | Training method, using method, device and equipment of machine learning model |
| CN113435998A (en) * | 2021-06-23 | 2021-09-24 | 平安科技（深圳）有限公司 | Loan overdue prediction method and device, electronic equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107622086A (en) * | 2017-08-16 | 2018-01-23 | 北京京东尚科信息技术有限公司 | A kind of clicking rate predictor method and device |
| CN110675243A (en) * | 2019-08-30 | 2020-01-10 | 北京银联金卡科技有限公司 | Machine learning-fused credit prediction overdue method and system |
-
2020
- 2020-04-13 CN CN202010283945.4A patent/CN111476658A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107622086A (en) * | 2017-08-16 | 2018-01-23 | 北京京东尚科信息技术有限公司 | A kind of clicking rate predictor method and device |
| CN110675243A (en) * | 2019-08-30 | 2020-01-10 | 北京银联金卡科技有限公司 | Machine learning-fused credit prediction overdue method and system |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112053225A (en) * | 2020-09-02 | 2020-12-08 | 中国银行股份有限公司 | Loan overdue prediction method and device, electronic equipment and storage medium |
| CN112150276A (en) * | 2020-09-30 | 2020-12-29 | 中国银行股份有限公司 | Training method, using method, device and equipment of machine learning model |
| CN113435998A (en) * | 2021-06-23 | 2021-09-24 | 平安科技（深圳）有限公司 | Loan overdue prediction method and device, electronic equipment and storage medium |
| CN113435998B (en) * | 2021-06-23 | 2023-05-02 | 平安科技（深圳）有限公司 | Loan overdue prediction method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111476658A (en) | Loan continuous overdue prediction method and device | |
| Fernando et al. | Development and validation of instruments adoption FinTech services in Indonesia (Perspective of trust and risk) | |
| CN112418653A (en) | Number portability and network diver identification system and method based on machine learning algorithm | |
| CN110704730A (en) | Product data pushing method and system based on big data and computer equipment | |
| CN112102073A (en) | Credit risk control method and system, electronic device and readable storage medium | |
| CN113609193A (en) | Method and device for training prediction model for predicting customer transaction behavior | |
| CN113052703A (en) | Transaction risk early warning method and device | |
| CN110796539A (en) | Credit investigation evaluation method and device | |
| CN111179055A (en) | Credit limit adjusting method and device and electronic equipment | |
| CN111091408A (en) | User identification model creating method and device and identification method and device | |
| CN114581249A (en) | Financial product recommendation method and system based on investment risk bearing capacity assessment | |
| CN112434862B (en) | Method and device for predicting financial dilemma of marketing enterprises | |
| CN111667307B (en) | Method and device for predicting financial product sales volume | |
| CN112132679A (en) | Product recommendation method and device | |
| CN111400413B (en) | Method and system for determining category of knowledge points in knowledge base | |
| CN114693428A (en) | Data determination method and device, computer readable storage medium and electronic equipment | |
| CN114170000A (en) | Credit card user risk category identification method, device, computer equipment and medium | |
| CN115545088A (en) | Model construction method, classification method and device and electronic equipment | |
| CN115860889A (en) | Financial loan big data management method and system based on artificial intelligence | |
| CN113849580A (en) | Subject rating prediction method and device, electronic equipment and storage medium | |
| CN114048330A (en) | Risk conduction probability knowledge graph generation method, device, equipment and storage medium | |
| CN113139842A (en) | Form processing method, device and system | |
| CN112634048A (en) | Anti-money laundering model training method and device | |
| CN111932310A (en) | Method and device for mining potential public customers of bank products | |
| CN111160929A (en) | Method and device for determining client type |
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 |