User repayment intention prediction method, device, computer equipment and storage medium

Abstract

The embodiments of the present application belong to the field of big data technology and are applied in the field of smart communities. They relate to a method, device, computer equipment and storage medium for predicting user repayment intentions, including: if the user loan data does not belong to standard data, performing removal operations and completion operations on the user loan data to obtain an initial user feature set; standardizing the initial user feature set to obtain target user feature data, performing key feature extraction operations and feature conversion operations on the target user feature data to obtain initial training data, calculating the proportion of financial data based on the target user feature data as proportion feature data, generating target training data based on the initial training data and the proportion feature data, training an initial overdue analysis model, and obtaining a target overdue analysis model; receiving data to be identified and inputting it into the target overdue analysis model to obtain the user's repayment intention. The target overdue analysis model can be stored in a blockchain. The present application improves the accuracy of predicting user repayment intentions.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and in particular to a method, device, computer equipment and storage medium for predicting user repayment intention. Background Art

[0002] The traditional risk control system has a long work chain and many manual approval links. Faced with complex, lengthy and cumbersome processes, the waiting period is longer. Because traditional risk control cannot analyze user behavior data and social data well, it is relatively biased towards enterprises and situations with larger amounts. When the amount of application data is larger, a large amount of manual labor will be invested in providing simple, repetitive and low-value labor, resulting in a large waste of manpower. Even after completing the risk control link, it is difficult to predict the consumer's willingness to repay. This is combined with the continuous development of micro-loan business for individuals in recent years. Preventing personal credit fraud and reducing the bad debt rate have become important considerations for conducting business.

[0003] Many companies have gradually begun to use important prediction models in the field of artificial intelligence, such as neural networks and decision trees, to predict possible overdue situations of users in loan scenarios. However, due to various reasons, such as poor training data, the prediction effect of the final prediction model is poor, making it difficult to accurately predict the user's repayment intention, and even more unable to help companies implement future strategies, adjust policies, and improve service quality. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose a method, device, computer equipment and storage medium for predicting user repayment intention, so as to improve the accuracy of predicting user repayment intention.

[0005] In order to solve the above technical problems, the present application embodiment provides a method for predicting user repayment intention, which adopts the following technical solution:

[0006] A method for predicting user repayment intention comprises the following steps:

[0007] receiving a user loan feature set, checking whether user loan data in the user loan feature set belongs to standard data, and if not, performing a removal operation and / or a completion operation on the user loan data to obtain an initial user feature set;

[0008] Performing standardization processing on the initial user feature data of the initial user feature set to obtain target user feature data, and performing a key feature extraction operation on the target user feature data to obtain key feature data;

[0009] Performing feature conversion operations on the key feature data based on a pre-configured mapping relationship to obtain initial training data, and calculating the proportions of multiple pre-specified financial data based on the target user feature data as proportion feature data, generating target training data based on the initial training data and the proportion feature data, and training a preset initial overdue analysis model through the target training data to obtain a target overdue analysis model;

[0010] Receive the data to be identified, input the data to be identified into the target overdue analysis model, and obtain the output of the user's repayment intention.

[0011] Furthermore, the step of checking whether the user loan data of the user loan feature set belongs to standard data, and if not, performing a removal operation and / or a completion operation on the user loan data includes:

[0012] Check whether the user loan data is dirty data;

[0013] If yes, it is determined that the user loan data does not belong to standard data, and the corresponding user loan data is removed;

[0014] If not, determine whether fields other than key fields in the user loan data are missing. If fields other than key fields are missing, determine that the user loan data does not belong to standard data, and fill in the preset default value to the position corresponding to the user loan data to obtain the initial user feature set.

[0015] Furthermore, the step of checking whether the user loan data is dirty data includes:

[0016] By calling a plurality of preset regular expressions, validation operations are performed on the user loan data respectively to check whether the user loan data is dirty data.

[0017] Furthermore, the target user feature data includes basic data, primary data and secondary data, and the step of performing a key feature extraction operation on the target user feature data to obtain the key feature data includes:

[0018] Converting the target user feature data into a user heat map based on a heat map generation function;

[0019] Determine, through the user heat map, secondary data having a correlation with the primary data greater than a correlation threshold as correlation data;

[0020] The correlation data, the master data and the basic data are used as the key feature data.

[0021] Furthermore, the initial user characteristic data includes a plurality of user data tables, and the step of performing standardization processing on the initial user characteristic data to obtain target user characteristic data includes:

[0022] The multiple user data tables are screened based on preset data table names to obtain an initial data table, and the initial data tables are merged to obtain the target user feature data.

[0023] Furthermore, the step of performing a feature conversion operation on the key feature data based on a pre-configured mapping relationship to obtain initial training data includes:

[0024] Identify text features in the key feature data, and convert the text features into numerical data based on a preset mapping relationship to obtain converted data;

[0025] The numerical features in the key feature data and the conversion data are used as the initial training data.

[0026] Furthermore, after the step of performing a key feature extraction operation on the target user feature data to obtain key feature data, the method further includes:

[0027] The date data in the target user characteristic data is parsed and converted to obtain target date data, and a user repayment cycle trend chart is generated based on the target date data.

[0028] In order to solve the above technical problems, the embodiment of the present application further provides a user repayment intention prediction device, which adopts the following technical solution:

[0029] A device for predicting user repayment intention, comprising:

[0030] A checking module, configured to receive a user loan feature set, check whether the user loan data in the user loan feature set belongs to standard data, and if not, perform a removal operation and / or a completion operation on the user loan data to obtain an initial user feature set;

[0031] An extraction module, used for performing standardization processing on the initial user feature data of the initial user feature set to obtain target user feature data, and performing a key feature extraction operation on the target user feature data to obtain key feature data;

[0032] a calculation module, configured to perform a feature conversion operation on the key feature data based on a pre-configured mapping relationship to obtain initial training data, and calculate the proportion of multiple pre-specified financial data based on the target user feature data as proportion feature data, generate target training data based on the initial training data and the proportion feature data, and train a preset initial overdue analysis model through the target training data to obtain a target overdue analysis model; and

[0033] The acquisition module is used to receive the data to be identified, input the data to be identified into the target overdue analysis model, and obtain the output of the user's repayment intention.

[0034] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0035] A computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the above-mentioned user repayment intention prediction method when executing the computer-readable instructions.

[0036] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0037] A computer-readable storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by a processor, implement the steps of the above-mentioned method for predicting user repayment intention.

[0038] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0039] This application obtains an initial user feature set without missing data by completing or removing the user loan data in the user loan feature set. By standardizing the initial user feature set and extracting key features, data screening of the initial user feature set is achieved, and more representative key data is extracted; by performing feature conversion operations on key feature data, it is beneficial for the model to better identify and learn key feature data, and by adding proportion feature data to the target training data, the model can learn more proportional data, thereby improving the accuracy of model predictions. This application can predict the repayment intention of users for each repayment period through the target overdue analysis model, so as to promptly remind the company to take targeted measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0042] Figure 2 is a flowchart of an embodiment of a method for predicting user repayment intention according to the present application;

[0043] Figure 3 is a structural schematic diagram of an embodiment of a device for predicting user repayment intention according to the present application;

[0044] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application.

[0045] Figure numerals: 200, computer device; 201, memory; 202, processor; 203, network interface; 300, user repayment intention prediction device; 301, inspection module; 302, extraction module; 303, calculation module; 304, acquisition module. DETAILED DESCRIPTION

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0047] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0048] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0049] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0050] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0051] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.

[0052] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0053] It should be noted that the user repayment intention prediction method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the user repayment intention prediction device is generally set in the server / terminal device.

[0054] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0055] With the continuous development of micro-loan business for individuals in recent years, preventing personal credit fraud and reducing bad debt rates are the primary goals of conducting related businesses. Therefore, it is very important to use big data, artificial intelligence, and machine learning related technologies to analyze and predict the existing data within each enterprise, help financial institutions accurately assess personal credit status, and further improve credit risk prevention capabilities.

[0056] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for predicting user repayment intention according to the present application. The method for predicting user repayment intention comprises the following steps:

[0057] S1: receiving a user loan feature set, checking whether the user loan data in the user loan feature set belongs to standard data, and if not, performing a removal operation and / or a completion operation on the user loan data to obtain an initial user feature set.

[0058] In this embodiment, if so, the user loan feature set is directly used as the initial user feature set. The user loan data of this application corresponds to the user ID in a one-to-one relationship. The user loan data includes user feature information such as the user's basic information, bank flow, credit card bill, and user browsing behavior data. Among them, the user's basic information includes: spouse information, education information, occupation information, and personal property status. The user's bank flow includes: transaction type, transaction amount, and source of income. The user's credit card bill includes: bank identification, last bill amount, last repayment amount, adjustment amount, interest, available balance, cash advance amount, repayment status, credit card limit, current bill balance, current bill minimum repayment amount, number of consumption items, and current bill amount. By performing a completion operation and / or removal operation on the user loan data in the user loan feature set, an initial user feature set consisting of user loan data that all belong to standard data is obtained, thereby improving the quality of the initial user feature set.

[0059] In this embodiment, the electronic device (eg Figure 1 The server / terminal device shown in the figure can receive the user loan feature set through a wired connection or a wireless connection. It should be noted that the above wireless connection method may include but is not limited to 3G / 4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0060] Specifically, in step S1, the step of checking whether the user loan data of the user loan feature set belongs to standard data, and if not, performing a removal operation and / or a completion operation on the user loan data includes:

[0061] Check whether the user loan data is dirty data;

[0062] If yes, it is determined that the user loan data does not belong to standard data, and the corresponding user loan data is removed;

[0063] If not, determine whether fields other than key fields in the user loan data are missing. If fields other than key fields are missing, determine that the user loan data does not belong to standard data, and fill in the preset default value to the position corresponding to the user loan data to obtain the initial user feature set.

[0064] In this embodiment, after acquiring the user loan feature set, the application performs a random check to see if the user loan data in the user loan feature set is dirty data. If so, the corresponding user loan data is removed. If not, it is determined whether the fields other than the key fields in the user loan data are missing. If the fields other than the key fields are missing, it is determined that the user loan data does not belong to standard data, and the default values ​​are filled in according to the specific business data to the position corresponding to the user loan data to obtain the initial user feature set. Dirty data judgment includes: missing key fields, where the key fields include user name, gender, bank flow records, user browsing behavior, timestamps, and credit card bill records. If missing, they are marked as -1. The amount field is non-numeric. The country field is not in the pre-listed enumeration. Abnormal situations such as the timestamp field being non-date are all dirty data. The specific judgment process uses regular formulas for judgment.

[0065] The step of checking whether the user loan data is dirty data includes:

[0066] By calling a plurality of preset regular expressions, validation operations are performed on the user loan data respectively to check whether the user loan data is dirty data.

[0067] In this embodiment, for example, regular expression 1: ^-[country 1, country 2, country 3 ... country n]\d*|0$, where regular expression 1 matches the characters listed in [], and is used to determine whether the international field is in the pre-listed enumeration.

[0068] S2: performing standardization processing on the initial user feature data of the initial user feature set to obtain target user feature data, and performing a key feature extraction operation on the target user feature data to obtain key feature data.

[0069] In this embodiment, the data is standardized so as to achieve better results in subsequent data training, and key data features are extracted from the processed data to implement data screening of the initial user feature set and extract more representative key data.

[0070] Specifically, the initial user characteristic data includes a plurality of user data tables. In step S2, that is, the step of performing standardization processing on the initial user characteristic data to obtain the target user characteristic data includes:

[0071] The multiple user data tables are screened based on preset data table names to obtain an initial data table, and the initial data tables are merged to obtain the target user feature data.

[0072] In this embodiment, it can be known from the user ID that not every user has bank flow records, credit card bills and other information. At the same time, from the detailed data analysis in the table, it can be known that not every user has very complete data records. Therefore, this application takes the user as the dimension to take the records in the corresponding 5 tables to form a complete table. Filter the data before the time dimension and the loan schedule. The target data table example is as follows:

[0073]

[0074]

[0075] In addition, the target user feature data includes basic data, primary data and secondary data. In step S2, that is, the key feature extraction operation is performed on the target user feature data, the step of obtaining the key feature data includes:

[0076] Converting the target user feature data into a user heat map based on a heat map generation function;

[0077] Determine, through the user heat map, secondary data having a correlation with the primary data greater than a correlation threshold as correlation data;

[0078] The correlation data, the master data and the basic data are used as the key feature data.

[0079] In this embodiment, data visualization: import the target user feature data into the seaborn library (visualization library), and generate a heat map through the heat map generation function (heatmap) of the seaborn library to make the data more intuitive. Seaborn is a more advanced API encapsulation based on matplotlib, which makes it easier to draw pictures. In most cases, seaborn can be used to make very representative pictures.

[0080] In this application, the primary data is the data that is most closely related to whether the user can successfully repay the loan on time in the loan scenario of this application. The primary data is specified by the back-end personnel according to the changes in the scenario, and the computer determines the type of the primary data. Among them, the secondary data is the data of the same type as the primary data. For example, if the type of the primary data is amount data, then the corresponding other amount data is secondary data.

[0081] The specific code example of the heat map generation process is as follows:

[0082] #Using seaborn library

[0083] importseaborn as sns

[0084] #Correlation result data table

[0085] corrmat=data_train[internal_chars].corr()

[0086] #Heatmap

[0087] sns.heatmap(corrmat, square=False, linewidths=.5, annot=True).

[0088] An example of the generated heat map is shown in the figure below, which realizes data visualization:

[0089]

[0090] Among them, the primary data is the current period's bill balance (type is amount category), and the secondary data is other amount category data, such as credit card limit, previous period repayment amount, previous period bill amount, current period bill amount, etc.

[0091] According to the heat map, the correlation analysis of the main data and the secondary data is performed, and the secondary data is selected to obtain the correlation data. For example, in the heat map, the correlation coefficient between "this period's bill amount" and "this period's bill balance" is 0.81, which has a large correlation. The correlation coefficient between "credit card limit" and "previous period repayment amount" is 0.51; the correlation coefficient between "credit card limit" and "previous period bill amount" is 0.55; the correlation coefficient between "credit card limit" and "current period bill balance" is 0.5; the correlation coefficient between "credit card limit" and "current period bill amount" is 0.5; there is correlation, but the correlation is not high, that is, the influence is general. The correlation coefficient between "previous period repayment amount" and "previous period bill amount" is 0.71, which has a large correlation. The correlation coefficient between "previous period repayment amount" and "current period bill balance" is 0.12; the correlation coefficient between "previous period repayment amount" and "current period bill amount" is 0.25; basically presents a non-correlated state. By analogy, the final selection and the main data (i.e., the current period's bill balance) have a highly collinear characteristic: the current period's bill amount and the previous period's repayment amount are used as the correlation data.

[0092] S3: Perform feature conversion operations on the key feature data based on a preconfigured mapping relationship to obtain initial training data, and calculate the proportion of multiple pre-specified financial data based on the target user feature data as proportion feature data, generate target training data based on the initial training data and the proportion feature data, and train a preset initial overdue analysis model through the target training data to obtain a target overdue analysis model.

[0093] In this embodiment, the key feature data is subjected to a feature conversion operation, that is, it is re-encoded. In this application, the proportion of multiple pre-specified financial data is calculated based on the target user feature data, and the step of calculating the proportion feature data specifically includes: calculating the proportion of various debit cards, the proportion of various credit cards, the proportion of various bank cards, the proportion of failed repayments, the proportion of failed loan applications, etc. based on the target user feature data as the proportion feature data. In the training data, the overdue mark is the label of the training data. The initial expected analysis model of this application adopts random forest and operates based on the sklearn library. sklearn integrates a large number of knowledge learning algorithms and can be used for knowledge learning. The specific code examples are as follows:

[0094] "from sklearn.ensemble import RandomForestClassifier as RFC

[0095] from sklearn.model_selection import GridSearchCV”

[0096] This application obtains the final desired prediction model, namely the target overdue analysis model, by training and verifying the initial overdue analysis model.

[0097] Specifically, in step S3, that is, the step of performing a feature conversion operation on the key feature data based on the pre-configured mapping relationship to obtain initial training data includes:

[0098] Identify text features in the key feature data, and convert the text features into numerical data based on a preset mapping relationship to obtain converted data;

[0099] The numerical features in the key feature data and the conversion data are used as the initial training data.

[0100] In this embodiment, feature conversion is information digitization. There are plain text features and numerical features in the key feature data, and the numerical features include discrete numerical features and continuous numerical features. Numerical features can generally be used directly in model training, but sometimes continuous variables are discretized for the stability and robustness of the model. At the same time, the text features are converted into numerical data through a pre-configured mapping relationship to obtain converted data, which is convenient for subsequent model analysis. An example is shown below:

[0101] Table 1 Basic information of users

[0102] User ID gender Profession educate Spouse Place of residence String Eum Eum Eum Eum Eum

[0103] Among them, Eum represents enumeration, String represents string, and the specific text features in Table 1 are converted into conversion data: [1008611,1,5,3,1,4,2,2].

[0104] Table 2 User's bank statement

[0105]

[0106] Among them, Double represents floating-point data. The specific text features in Table 2 are converted into conversion data: [1008611, 0, 289.9000, 1, 1631776833475, 2].

[0107] Table 3 User browsing behavior

[0108] User ID Browsing behavior data Timestamp String String 1317475766833

[0109] The converted data is: [1008611,202104091532432800001,1317475766833].

[0110] Table 4 User's credit card bill

[0111] User ID String Timestamp 1317475766833 Bank logo Eum Last bill amount Double Adjustment amount Double Interest Double Available balance Double Cash advance amount Double Repayment status Eum Credit card limit Double Current bill balance Double Minimum payment for this bill Double Number of consumption transactions Int Amount of current bill Double

[0112] Table 5 User's bank statement

[0113] User ID Overdue mark Overdue days String Eum Eum

[0114] The converted data is: [1008611,1,10].

[0115] Among them, String, Eum, and Double all represent data types, String is a string, Eum is an enumeration, and Double is floating-point data.

[0116] S4: receiving the data to be identified, inputting the data to be identified into the target overdue analysis model, and obtaining the output of the user's repayment intention.

[0117] In this embodiment, the target overdue analysis model is applied to the user data to be repaid in the current month (i.e., the data to be identified) to predict the user's repayment intention. The user's repayment intention is output, and the user's repayment intention includes yes and no. Among them, "yes" means that the user will repay or not repay on time, and "no" means that the user will not repay on time.

[0118] After obtaining the output of the user's repayment intention, it also includes: judging whether the user's repayment intention is overdue repayment. If the user's repayment intention is overdue repayment, the user's contact information is searched according to the user ID, and a repayment reminder is sent to the user terminal through the user's contact information. At the same time, a reminder notification that the user may be overdue is sent to the enterprise. The enterprise can then take targeted measures, such as reducing the loan amount for the customer, and the approval situation can also be slowed down. In this way, the capital flow of the enterprise's micro-loan business can be adjusted to reduce the bad debt rate by allowing the funds to flow reasonably. At the same time, it can also effectively prevent personal credit fraud and reduce the bad debt rate.

[0119] In some optional implementations of this embodiment, after step S2, the electronic device may further perform the following steps: performing parsing and conversion operations on the date data in the target user characteristic data to obtain target date data, and generating a user repayment cycle trend chart based on the target date data.

[0120] Use existing tools to quickly parse date data for visualization and repayment cycle trend display. The specific steps for quickly parsing date data include: There is date data to be analyzed in the data set, but the type may be String, or inconsistent date formats such as %Y-%m-%d, %d-%m-%Y, etc. Establish a two-dimensional coordinate system, use the target date data as the horizontal axis, and the corresponding target user feature data The amount of each period of the user's repayment is used as the vertical axis to generate the user's repayment cycle trend chart.

[0121] Specifically, the step of performing parsing and conversion operations on the date data in the intermediate user characteristic data includes:

[0122] The pandas tool is called to parse the date data to obtain a date value, and the target date data is generated based on the date value and a preset date template.

[0123] In this embodiment, there is date data to be analyzed in the data set, but the type may be String, or inconsistent date formats such as %Y-%m-%d, %d-%m-%Y, etc. This application uses pandas to parse these date data, change the type of date data to datetime, and unify it into date format (here formatted as "%m / %d / %y"). Pandas is used for data manipulation and analysis.

[0124] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned target overdue analysis model, the above-mentioned target overdue analysis model can also be stored in a node of a blockchain.

[0125] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0126] This application can be applied in the field of smart communities, thereby promoting the construction of smart cities.

[0127] This application obtains an initial user feature set without missing data by completing or removing the user loan data in the user loan feature set. By standardizing the initial user feature set and extracting key features, data screening of the initial user feature set is achieved, and more representative key data is extracted; by performing feature conversion operations on key feature data, it is beneficial for the model to better identify and learn key feature data, and by adding proportion feature data to the target training data, the model can learn more proportional data, thereby improving the accuracy of model predictions. This application can predict the repayment intention of users for each repayment period through the target overdue analysis model, so as to promptly remind the company to take targeted measures.

[0128] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the computer-readable instructions are executed, they can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0129] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0130] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a device for predicting user repayment intention. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0131] like Figure 3 As shown, the user repayment intention prediction device 300 described in this embodiment includes: a checking module 301, an extraction module 302, a calculation module 303 and an acquisition module 304. Among them: the checking module 301 is used to receive a user loan feature set, check whether the user loan data of the user loan feature set belongs to standard data, and if not, perform a removal operation and / or a completion operation on the user loan data to obtain an initial user feature set; the extraction module 302 is used to perform standardization processing on the initial user feature data of the initial user feature set to obtain target user feature data, and perform a key feature extraction operation on the target user feature data to obtain key feature data; the calculation module 303 is used to perform a feature conversion operation on the key feature data based on a pre-configured mapping relationship to obtain initial training data, and calculate the proportion of multiple pre-specified financial data based on the target user feature data as proportion feature data, generate target training data based on the initial training data and the proportion feature data, and train a preset initial overdue analysis model through the target training data to obtain a target overdue analysis model; and the obtaining module 304 is used to receive data to be identified, input the data to be identified into the target overdue analysis model, and obtain the output user repayment intention.

[0132] In this embodiment, the present application obtains an initial user feature set without missing data by completing or removing the user loan data in the user loan feature set. By standardizing the initial user feature set and extracting key features, data screening of the initial user feature set is achieved, and more representative key data is extracted; by performing feature conversion operations on the key feature data, it is beneficial for the model to better identify and learn the key feature data, and by adding proportion feature data to the target training data, the model can learn more proportional data, thereby improving the accuracy of the model prediction. The present application can predict the repayment intention of users for each repayment period through the target overdue analysis model, so as to promptly remind the enterprise to take targeted measures.

[0133] The checking module 301 includes a checking submodule, a removing submodule and a completing submodule, wherein the checking submodule is used to check whether the user loan data is dirty data; the removing submodule is used to determine that the user loan data does not belong to standard data and remove the corresponding user loan data when the user loan data is dirty data; the completing submodule is used to determine whether fields other than key fields in the user loan data are missing when the user loan data does not belong to dirty data; if fields other than key fields are missing, it is determined that the user loan data does not belong to standard data, and the preset default value is completed to the position corresponding to the user loan data to obtain the initial user feature set.

[0134] In some optional implementations of this embodiment, the determination submodule is further used to: perform validation operations on the user loan data respectively by calling a plurality of preset regular expressions to check whether the user loan data is dirty data.

[0135] In some optional implementations of this embodiment, the extraction module 302 is further used to: filter the multiple user data tables based on preset data table names to obtain initial data tables, and merge the initial data tables to obtain the target user feature data.

[0136] The target user characteristic data includes basic data, main data and secondary data, and the extraction module 302 includes a conversion submodule, a correlation data generation submodule and a key characteristic data generation submodule. The conversion submodule is used to convert the target user characteristic data into a user heat map based on a heat map generation function; the correlation data generation submodule is used to determine the secondary data with a correlation greater than a correlation threshold with the main data through the user heat map as the correlation data; the key characteristic data generation submodule is used to use the correlation data, main data and basic data as the key characteristic data.

[0137] The calculation module 303 includes an identification submodule and an initial training data generation submodule, wherein the identification submodule is used to identify text features in the key feature data, convert the text features into numerical data based on a preset mapping relationship, and obtain conversion data; the initial training data generation submodule is used to use the numerical features in the key feature data and the conversion data as the initial training data.

[0138] In some optional implementations of this embodiment, the above-mentioned device 300 also includes: a parsing module, which is used to perform parsing and conversion operations on the date data in the target user characteristic data, obtain target date data, and generate a user repayment cycle trend chart based on the target date data.

[0139] In some optional implementations of this embodiment, the above-mentioned parsing module is further used to: call a pandas tool to parse the date data, obtain a date value, and generate the target date data based on the date value and a preset date template.

[0140] This application obtains an initial user feature set without missing data by completing or removing the user loan data in the user loan feature set. By standardizing the initial user feature set and extracting key features, data screening of the initial user feature set is achieved, and more representative key data is extracted; by performing feature conversion operations on key feature data, it is beneficial for the model to better identify and learn key feature data, and by adding proportion feature data to the target training data, the model can learn more proportional data, thereby improving the accuracy of model predictions. This application can predict the repayment intention of users for each repayment period through the target overdue analysis model, so as to promptly remind the company to take targeted measures.

[0141] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0142] The computer device 200 includes a memory 201, a processor 202, and a network interface 203 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 200 having components 201-203, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field-programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0143] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0144] The memory 201 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 201 can be an internal storage unit of the computer device 200, such as a hard disk or memory of the computer device 200. In other embodiments, the memory 201 can also be an external storage device of the computer device 200, such as a plug-in hard disk equipped on the computer device 200, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory 201 can also include both the internal storage unit of the computer device 200 and its external storage device. In this embodiment, the memory 201 is generally used to store the operating system and various application software installed on the computer device 200, such as computer-readable instructions of the user repayment intention prediction method. In addition, the memory 201 can also be used to temporarily store various data that have been output or are to be output.

[0145] The processor 202 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 202 is generally used to control the overall operation of the computer device 200. In this embodiment, the processor 202 is used to run computer-readable instructions or process data stored in the memory 201, such as computer-readable instructions for running the user repayment intention prediction method.

[0146] The network interface 203 may include a wireless network interface or a wired network interface. The network interface 203 is generally used to establish a communication connection between the computer device 200 and other electronic devices.

[0147] In this embodiment, by performing feature conversion operations on key feature data, the model is helped to better identify and learn the key feature data, and proportion feature data is added to the target training data so that the model can learn more proportional data, thereby improving the accuracy of model prediction.

[0148] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the user repayment intention prediction method as described above.

[0149] In this embodiment, by performing feature conversion operations on key feature data, the model is helped to better identify and learn the key feature data, and proportion feature data is added to the target training data so that the model can learn more proportional data, thereby improving the accuracy of model prediction.

[0150] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0151] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for predicting user repayment intention, characterized in that: The steps include: receiving a user loan feature set, checking whether user loan data in the user loan feature set belongs to standard data, and if not, performing a removal operation and / or a completion operation on the user loan data to obtain an initial user feature set; Standardizing the initial user feature data of the initial user feature set to obtain target user feature data, and performing a key feature extraction operation on the target user feature data to obtain key feature data, wherein the target user feature data includes basic data, primary data, and secondary data; Performing a feature conversion operation on the key feature data based on a pre-configured mapping relationship to obtain initial training data, and calculating the proportion of multiple pre-specified financial data based on the target user feature data as proportion feature data, generating target training data based on the initial training data and the proportion feature data, and training a preset initial overdue analysis model through the target training data to obtain a target overdue analysis model; Receiving data to be identified, inputting the data to be identified into the target overdue analysis model, and obtaining an output of the user's repayment intention; The step of performing a key feature extraction operation on the target user feature data to obtain the key feature data specifically includes: Converting the target user feature data into a user heat map based on a heat map generation function; Determine, through the user heat map, secondary data having a correlation with the primary data greater than a correlation threshold as correlation data; Using the correlation data, the master data and the basic data as the key feature data; The step of calculating the proportion of multiple pre-specified financial data based on the target user characteristic data as the proportion characteristic data specifically includes: Based on the target user characteristic data, the proportion of various debit cards, the proportion of various credit cards, the proportion of various bank cards, the proportion of failed repayments, and the proportion of failed loan applications are calculated as proportion characteristic data.

2. The method for predicting user repayment intention according to claim 1, characterized in that: The step of checking whether the user loan data of the user loan feature set belongs to standard data, and if not, performing a removal operation and / or a completion operation on the user loan data comprises: Check whether the user loan data is dirty data; If yes, it is determined that the user loan data does not belong to standard data, and the corresponding user loan data is removed; If not, determine whether fields other than key fields in the user loan data are missing. If fields other than key fields are missing, determine that the user loan data does not belong to standard data, and fill in the preset default value to the position corresponding to the user loan data to obtain the initial user feature set.

3. The method for predicting user repayment intention according to claim 2, characterized in that: The step of checking whether the user loan data is dirty data comprises: By calling a plurality of preset regular expressions, validation operations are performed on the user loan data respectively to check whether the user loan data is dirty data.

4. The method for predicting user repayment intention according to claim 1, characterized in that: The initial user characteristic data includes a plurality of user data tables, and the step of performing standardization processing on the initial user characteristic data to obtain target user characteristic data includes: The multiple user data tables are screened based on preset data table names to obtain an initial data table, and the initial data tables are merged to obtain the target user feature data.

5. The method for predicting user repayment intention according to claim 1, characterized in that: The step of performing a feature conversion operation on the key feature data based on a pre-configured mapping relationship to obtain initial training data comprises: Identify text features in the key feature data, and convert the text features into numerical data based on a preset mapping relationship to obtain converted data; The numerical features in the key feature data and the conversion data are used as the initial training data.

6. The method for predicting user repayment intention according to claim 1, characterized in that: After the step of performing a key feature extraction operation on the target user feature data to obtain key feature data, the method further includes: The date data in the target user characteristic data is parsed and converted to obtain target date data, and a user repayment cycle trend chart is generated based on the target date data.

7. A device for predicting user repayment intention, characterized in that: include: A checking module, configured to receive a user loan feature set, check whether the user loan data in the user loan feature set belongs to standard data, and if not, perform a removal operation and / or a completion operation on the user loan data to obtain an initial user feature set; An extraction module, used to perform standardization processing on the initial user feature data of the initial user feature set to obtain target user feature data, and perform key feature extraction operation on the target user feature data to obtain key feature data, wherein the target user feature data includes basic data, primary data and secondary data; The extraction module includes: a conversion submodule, a correlation data generation submodule and a key feature data generation submodule: The conversion submodule is used to convert the target user feature data into a user heat map based on a heat map generation function; The correlation data generating submodule is used to determine, through the user heat map, secondary data having a correlation with the primary data greater than a correlation threshold, as correlation data; The key feature data generating submodule is used to use the correlation data, the master data and the basic data as the key feature data; A calculation module, configured to perform a feature conversion operation on the key feature data based on a pre-configured mapping relationship to obtain initial training data, and calculate the proportion of multiple pre-specified financial data based on the target user feature data as proportion feature data, generate target training data based on the initial training data and the proportion feature data, and train a preset initial overdue analysis model through the target training data to obtain a target overdue analysis model; The calculation module is further used to calculate the proportion of various debit cards, the proportion of various credit cards, the proportion of various bank cards, the proportion of failed repayments, and the proportion of failed loan applications based on the target user characteristic data as proportion characteristic data; and The acquisition module is used to receive the data to be identified, input the data to be identified into the target overdue analysis model, and obtain the output of the user's repayment intention.

8. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the method for predicting user repayment intention as described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the method for predicting user repayment intention as described in any one of claims 1 to 6.

Images