An intelligent prediction method and system for income capacity based on credit behavior characteristics

Abstract

The present invention relates to an intelligent prediction method and system for income capacity based on credit behavior characteristics, the method comprising: obtaining customer credit performance, customer basic information and macro-indicator data, extracting credit behavior characteristics, and establishing a feature data set; taking income level as a prediction target, establishing an XGBoost prediction model for income level prediction, and using training set data to train the XGBoost prediction model; for multi-classification income level prediction problems, taking the same-level accuracy rate as an optimization target, and adjusting model parameters. The present invention extracts credit behavior characteristics, uses a gradient boosting tree XGBoost algorithm to establish a prediction model, and adjusts model parameters with the same-level accuracy rate as an optimization target, thereby achieving accurate prediction of customer income levels, providing effective tools for credit approval and customer management, reducing data costs, and improving customer portraits.

Description

Technical Field

[0001] The present invention relates to the field of financial credit and artificial intelligence technology, and in particular to a method and system for intelligently predicting income capacity based on credit behavior characteristics. Background Art

[0002] In the field of financial credit, the income capacity of customers is an important basis for assessing the credit status of customers, determining credit limits, interest rates, and conducting customer marketing. Traditional income capacity assessment methods mainly rely on official data such as provident fund payment records and social security payment status, as well as factual or predicted income level data provided by external data source vendors. However, with the changes in the market environment and the increasing difficulty of data acquisition, these methods have gradually exposed many limitations and cannot meet the growing refinement and real-time needs of credit business. At present, the existing income capacity prediction methods have the following defects:

[0003] Issues with data coverage and accuracy: Housing fund and social security data are missing and updated in a untimely manner, while external data sources face challenges such as high costs and low coverage, resulting in a weak data foundation that makes it difficult to support accurate income forecasts.

[0004] Insufficient feature utilization: Existing methods often only focus on the basic information of customers and some credit performance, ignoring important features that can reflect income levels, such as customer occupations and consumption habits, resulting in limited model prediction capabilities.

[0005] Improper balance between model selection and complexity: Traditional statistical models cannot effectively capture complex patterns in the data, while complex models such as neural networks are prone to overfitting, making it difficult to find a balance between model complexity and generalization ability.

[0006] Insufficient consideration of external factors: Customer revenue is affected by multiple external factors such as the macroeconomic environment, policy changes, and industry development, but existing models often fail to fully incorporate these factors, causing the forecast results to deviate from the actual situation. Summary of the invention

[0007] In view of this, it is necessary to provide an intelligent prediction method and system for income capacity based on credit behavior characteristics to solve the above-mentioned defects of the prior art.

[0008] In order to solve the above problems, in a first aspect, an embodiment of the present invention provides an intelligent prediction method of income capacity based on credit behavior characteristics, comprising:

[0009] Obtain customer credit performance, basic customer information and macro-indicator data, extract credit behavior characteristics, and establish a feature data set; divide the feature data set into a training set and a test set;

[0010] Taking income level as the prediction target, an XGBoost prediction model for income level prediction is established, and the XGBoost prediction model is trained using the training set data;

[0011] For the problem of multi-classification income level prediction, the same-level accuracy is taken as the optimization target to adjust the model parameters of the XGBoost prediction model.

[0012] Preferably, the method further comprises:

[0013] Input the test set data into the trained XGBoost prediction model to obtain the income level prediction results;

[0014] According to the predicted income level and the actual income level, the accuracy of the same level, the accuracy of the upper and lower adjacent levels, and the accuracy of the upper and lower adjacent levels are calculated to evaluate the accuracy of the model;

[0015] Establish an efficient model deployment and update process to ensure that the model can be put into production environment in a timely manner.

[0016] Preferably, after obtaining the customer's credit performance, basic customer information and macro-indicator data, the method further comprises:

[0017] Clean the data and handle missing values, outliers, and inconsistent data.

[0018] Preferably, the extracting of credit behavior characteristics includes:

[0019] Credit behavior characteristics are extracted based on customer credit performance, basic customer information and macro-indicator data; among which, credit behavior characteristics include the customer's region, the nature of the customer's unit, occupational information and educational performance.

[0020] Preferably, before establishing an XGBoost prediction model for income level prediction with income level as the prediction target, the method further includes:

[0021] Assess the accuracy and coverage of different data sources in describing income levels and select the best data source;

[0022] Income brackets were set based on sample averages.

[0023] Preferably, the method of adjusting the model parameters of the XGBoost prediction model by taking the same gear accuracy as the optimization target specifically includes:

[0024] Through the Optuna parameter adjustment method, the hyperparameters of the XGBoost prediction model are tuned with the same-level accuracy as the optimization goal to obtain the optimal model parameter combination; among them, the hyperparameters of the XGBoost prediction model include tree depth, learning rate, and regularization coefficient.

[0025] In a second aspect, an embodiment of the present invention provides an intelligent prediction system for income capacity based on credit behavior characteristics, including:

[0026] Feature engineering module, used to obtain customer credit performance, basic customer information and macro-indicator data, extract credit behavior characteristics, and establish feature data sets; divide the feature data sets into training sets and test sets;

[0027] A model training module is used to establish an XGBoost prediction model for income level prediction with income level as the prediction target, and train the XGBoost prediction model using the training set data;

[0028] The model optimization module is used to adjust the model parameters of the XGBoost prediction model for the multi-classification income level prediction problem, taking the same-level accuracy as the optimization target.

[0029] In a third aspect, the present invention further provides an electronic device, including a memory and a processor, wherein:

[0030] The memory is used to store programs;

[0031] The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the method for intelligently predicting income capacity based on credit behavior characteristics as described in the embodiment of the first aspect of the present invention.

[0032] In a fourth aspect, the present invention further provides a computer-readable storage medium for storing computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the method for intelligently predicting income capacity based on credit behavior characteristics as described in the embodiment of the first aspect of the present invention.

[0033] Compared with the prior art, the income capacity intelligent prediction method and system based on credit behavior characteristics provided by the present invention has the following beneficial effects:

[0034] 1) The present invention significantly improves the accuracy of income level prediction through comprehensive evaluation and selection of data sources, in-depth mining and utilization of features, selection of appropriate models and optimization parameters, and other technical means, providing a more reliable decision-making basis for credit approval and customer management.

[0035] 2) Through the Optuna parameter adjustment method, the present invention enhances the generalization ability of the model, enables the model to better adapt to different scenarios and data changes, and improves the stability and practicality of the model.

[0036] 3) By comprehensively evaluating and selecting data sources, the reliance on external data sources is reduced, thereby reducing data costs and improving business efficiency. By deeply mining and utilizing features, the present invention further enriches customer portraits, provides support for personalized services and precision marketing, and improves customer experience and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A flow chart of the method for intelligently predicting income capacity based on credit behavior characteristics provided by the present invention;

[0038] Figure 2 A structural block diagram of the income capacity intelligent prediction system based on credit behavior characteristics provided by the present invention;

[0039] Figure 3 This is a structural block diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0040] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0041] 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.

[0042] At present, the existing income capacity prediction methods have the following defects: First, the data coverage is insufficient. The provident fund, social security and external data sources that are mainly relied on are not updated in time, are seriously missing or are costly, which limits the accuracy and wide application of the prediction. Secondly, the features are not fully utilized, and key information such as customer credit performance, occupation, consumption habits, etc. is not fully mined, resulting in limited model prediction capabilities. Furthermore, it is difficult to balance model selection and complexity. Traditional methods are difficult to capture complex patterns, and complex models are prone to overfitting, affecting the stability of predictions. Finally, external factors are not considered enough, such as macroeconomics, policy changes, etc., which are not fully incorporated into the model, resulting in prediction results that deviate from reality. These defects jointly restrict the effectiveness and practicality of existing income capacity prediction methods.

[0043] In view of this, the present invention proposes an intelligent prediction method for income capacity based on credit behavior characteristics to address the problems of data coverage and accuracy, insufficient feature utilization, improper balance between model selection and complexity, and insufficient consideration of external factors in existing income capacity prediction methods. This method comprehensively evaluates and selects the best data source, extracts credit behavior characteristics, uses the gradient boosting tree XGBoost algorithm to establish a prediction model, and adjusts the model parameters with the same level accuracy as the optimization goal, thereby achieving accurate prediction of customer income levels, providing effective tools for credit approval and customer management, reducing data costs, and improving customer portraits. The following will be explained and introduced through multiple embodiments.

[0044] Figure 1 The flow chart of the intelligent prediction method of income capacity based on credit behavior characteristics provided by the present invention is shown in FIG. Figure 1 The method for intelligently predicting income capacity based on credit behavior characteristics provided by the present invention at least comprises the following steps:

[0045] Step S1, obtain customer credit performance, customer basic information and macro indicator data, extract credit behavior characteristics, and establish a feature data set; divide the feature data set into a training set and a test set.

[0046] Specifically, the credit performance of a customer refers to the repayment ability, repayment willingness and other credit-related behavioral characteristics shown by the customer in the credit activities. These data usually come from the customer's credit record, including but not limited to loan history, repayment record, overdue situation, credit score, etc. The basic information of a customer refers to the data that can identify the customer's identity and describe the basic characteristics of the customer. This information usually includes the customer's name, gender, age, contact information, occupation, education, family status, etc. Basic information is the basis for understanding the customer's background and assessing the customer's potential risks, and is also an indispensable part of the credit approval process. Macro indicators refer to data that reflect macro factors such as the overall economic environment and industry trends. In the field of credit, macro indicators are of great significance in assessing the economic environment in which customers are located and predicting future repayment ability. Macro indicators include gross domestic product (GDP), inflation rate, unemployment rate, interest rate level, international trade balance, etc.

[0047] In a preferred embodiment of the present invention, after obtaining the customer credit performance, basic customer information and macro-indicator data, the method further includes: cleaning the data, processing missing values, outliers and inconsistent data. The purpose of data cleaning is to eliminate errors, anomalies and inconsistencies in the data to make the data more accurate and reliable. For missing values ​​in the data, appropriate strategies need to be adopted to fill or delete them. For example, statistical values ​​such as mean, median, mode, etc. can be used to fill missing values. Outliers refer to values ​​that are significantly deviated from the normal data range, which may be caused by data entry errors, system failures or special events. Outliers need to be identified and processed to avoid their adverse effects on model training. During the data integration process, it may be found that there are inconsistencies in the data between different data sources. For example, the customer's name, address or occupation information may be different in different data sources. In this case, data comparison and correction are required to ensure the consistency and accuracy of the data. Through data cleaning, errors and anomalies in the data can be eliminated, and the accuracy and reliability of the data can be improved, thereby laying a solid foundation for subsequent model training and prediction.

[0048] Furthermore, after obtaining customer credit performance, basic customer information and macro-indicator data and performing data cleaning, credit behavior characteristics can be further extracted.

[0049] Based on the customer's credit performance, basic customer information and macro-indicator data, we deeply explore the key characteristics related to the customer's income, namely the credit behavior characteristics. Among them, the credit behavior characteristics include the customer's region, the nature of the customer's unit, occupation information and educational performance.

[0050] It is understandable that in machine learning models, feature selection and extraction is one of the key factors affecting model performance. For income prediction problems, features extracted directly from raw data may not directly reflect income levels, or may have a low correlation with income levels. Therefore, it is necessary to deeply explore key features that are highly correlated with income levels to improve the prediction accuracy of the model. By extracting and utilizing these key features, the prediction accuracy and generalization ability of the model can be significantly improved.

[0051] Step S2, taking income level as the prediction target, establishing an XGBoost prediction model for income level prediction, and training the XGBoost prediction model using training set data.

[0052] In a preferred embodiment of the present invention, before establishing an XGBoost prediction model for income level prediction with income level as the prediction target, the data source may be pre-selected and the income level division may be set.

[0053] First, evaluate the accuracy and coverage of different data sources in describing income levels, and select the best data source. Then, set the income level based on the sample average. In this embodiment, comprehensively evaluate the accuracy and coverage of provident fund, social security and external data sources in describing income levels, and then select the best performing data source as the prediction basis for the income model. The selection criteria can be based on multiple factors, such as data reliability, availability, and correlation with income levels. After selecting the data source, then sort the sample data by income level and then divide it into several levels. Each level represents a specific range of income levels, which serves as the prediction target of the income model.

[0054] This embodiment can improve the prediction accuracy of the income model by comprehensively evaluating the accuracy and coverage of different data sources and selecting the best data source as the prediction basis.

[0055] Among them, XGBoost (eXtreme Gradient Boosting) is an efficient gradient boosting tree algorithm, which is used to process the multi-classification income level prediction problem in the embodiment of the present invention.

[0056] In this embodiment, income levels are divided into multiple categories (such as low income, middle income, high income, etc.), and the task of the model is to predict which income level the customer belongs to. This embodiment uses income level as the prediction target and uses the XGBoost algorithm to establish a prediction model. This model accepts features such as customer credit performance, customer basic information, macro indicators, etc. as input, and outputs the predicted income level. During the training process, the model learns the correlation between features and income levels, and adjusts parameters to optimize prediction performance. After training is completed, the model can be used to predict the income level of new customers.

[0057] Step S3, for the multi-classification income level prediction problem, the same-level accuracy is used as the optimization target to adjust the model parameters of the XGBoost prediction model.

[0058] Specifically, through the Optuna parameter adjustment method, the hyperparameters of the XGBoost prediction model are tuned with the same-level accuracy as the optimization target to obtain the optimal model parameter combination, thereby improving the generalization ability of the model. Among them, the hyperparameters of the XGBoost prediction model include tree depth, learning rate, and regularization coefficient.

[0059] Among them, Optuna is an open source library for hyperparameter optimization. It provides a variety of optimization algorithms (such as grid search, random search, Bayesian optimization, etc.) to automatically find the best hyperparameter combination of the model. In this invention, Optuna's powerful optimization capabilities are used to automatically adjust the hyperparameters of the model to achieve better prediction results.

[0060] In this embodiment, the same-level accuracy is used as the optimization target, and the same-level accuracy refers to the ratio of the income level predicted by the model to the actual income level. By continuously adjusting the model parameters, the model can achieve a higher level of same-level prediction accuracy, thereby approaching the optimal value of the prediction. This is usually achieved through an iterative training process, and each iteration adjusts the parameters according to the performance of the current model in order to obtain better prediction results.

[0061] In a preferred embodiment of the present invention, the income capacity intelligent prediction method based on credit behavior characteristics further includes:

[0062] Step S4, input the test set data into the trained XGBoost prediction model to obtain the income level prediction result; according to the income level prediction result and the actual income level, calculate the accuracy of the same gear, the accuracy of the upper and lower adjacent gears, and the accuracy of the upper and lower adjacent gears to evaluate the accuracy of the model.

[0063] Specifically, the test set data is input into the trained XGBoost prediction model. The model will predict an income level for each customer based on the input feature data, using the learned decision tree and parameters. These predictions will be compared with the actual income levels in the test set to evaluate the accuracy of the model.

[0064] The following is an explanation of the same gear accuracy, upper and lower adjacent gear accuracy, and upper and lower adjacent two gear accuracy:

[0065] Same-level accuracy: One of the main indicators for evaluating the model's prediction performance. It measures the proportion of income levels predicted by the model that are exactly the same as the actual income level. The higher the same-level accuracy, the more accurate the model's predictions.

[0066] Accuracy of adjacent levels: In addition to the same level accuracy, we also focus on the accuracy of the levels where the model prediction results are adjacent to the actual results. This is because in practical applications, even if the model does not accurately predict a specific income level, if the prediction results are adjacent to the actual results, it can still be considered that the model has a certain degree of prediction ability.

[0067] Accuracy of adjacent levels: This is a broader accuracy indicator that takes into account the proportion of predicted results that differ from the actual results by no more than two levels. This indicator can further evaluate the model's error tolerance in predicting income levels.

[0068] In this embodiment, by calculating the same gear accuracy, the upper and lower adjacent gear accuracy, and the upper and lower adjacent two gear accuracy, the prediction results of the model can be compared with the actual income level to evaluate the accuracy of the model. In addition to the accuracy index, other evaluation indicators (such as mean square error, mean absolute error, etc.) and model stability, generalization ability, etc. can also be combined to comprehensively evaluate the model. According to the evaluation results, we can optimize and adjust the model, such as adjusting feature selection, model parameters, etc., to improve the prediction accuracy of the model.

[0069] Step S5: Establish an efficient model deployment and update process to ensure that the model can be put into production environment in a timely manner.

[0070] Specifically, model deployment is a key step in the transition from theory to practice. It refers to the process of integrating a trained, verified, and optimized model into an actual application system so that it can play a role in actual business scenarios. As the business environment changes and data is updated, the performance of the model may be affected. Therefore, the model needs to be updated regularly to maintain its accuracy and effectiveness.

[0071] It should be noted that steps S1 to S5 described in the present invention are only an exemplary process description, which is used to show a possible way of implementing the present invention. However, this does not mean that these steps must be performed strictly in this order. In practical applications, according to specific data conditions, business needs and model performance requirements, the order of these steps can be flexibly adjusted without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

[0072] Compared with the prior art, the income capacity intelligent prediction method based on credit behavior characteristics provided by the present invention has the following beneficial effects:

[0073] 1) The present invention significantly improves the accuracy of income level prediction through comprehensive evaluation and selection of data sources, in-depth mining and utilization of features, selection of appropriate models and optimization parameters, and other technical means, providing a more reliable decision-making basis for credit approval and customer management.

[0074] 2) Through the Optuna parameter adjustment method, the present invention enhances the generalization ability of the model, enables the model to better adapt to different scenarios and data changes, and improves the stability and practicality of the model.

[0075] 3) By comprehensively evaluating and selecting data sources, the reliance on external data sources is reduced, thereby reducing data costs and improving business efficiency. By deeply mining and utilizing features, the present invention further enriches customer portraits, provides support for personalized services and precision marketing, and improves customer experience and satisfaction.

[0076] Figure 2 The structural block diagram of the income capacity intelligent prediction system based on credit behavior characteristics provided by the present invention is shown in FIG. Figure 2 , the income capacity intelligent prediction system 200 based on credit behavior characteristics includes:

[0077] The feature engineering module 201 is used to obtain customer credit performance, basic customer information and macro-indicator data, extract credit behavior characteristics, and establish a feature data set; divide the feature data set into a training set and a test set;

[0078] A model training module 202 is used to establish an XGBoost prediction model for income level prediction with income level as the prediction target, and train the XGBoost prediction model using training set data;

[0079] The model optimization module 203 is used to adjust the model parameters of the XGBoost prediction model for the multi-classification income level prediction problem, taking the same-level accuracy rate as the optimization target.

[0080] The intelligent prediction system for income capacity based on credit behavior characteristics provided by the present invention executes the intelligent prediction method for income capacity based on credit behavior characteristics provided by the aforementioned embodiments through the above-mentioned modules. The intelligent prediction method for income capacity based on credit behavior characteristics has been described in detail in the above-mentioned embodiments, and will not be repeated in this embodiment.

[0081] Figure 3 The structural block diagram of the electronic device provided by the present invention is as follows: Figure 3 As shown, the present invention also provides an electronic device, and the electronic device 300 can be a computing device such as a mobile terminal, a desktop computer, a notebook, a palm computer, and a server. The electronic device 300 includes a processor 301 and a memory 302, wherein the memory 302 stores an income capacity intelligent prediction program 303 based on credit behavior characteristics.

[0082] In some embodiments, the memory 302 may be an internal storage unit of a computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory 302 may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device. Furthermore, the memory 302 may also include both an internal storage unit of the computer device and an external storage device. The memory 302 is used to store application software and various types of data installed on the computer device, such as program codes installed on the computer device. The memory 302 may also be used to temporarily store data that has been output or is to be output. In one embodiment, when the income capacity intelligent prediction program 303 based on credit behavior characteristics is executed by the processor 301, the following steps are implemented:

[0083] Obtain customer credit performance, basic customer information and macro-indicator data, extract credit behavior characteristics, and establish a feature data set; divide the feature data set into a training set and a test set;

[0084] Taking income level as the prediction target, an XGBoost prediction model for income level prediction is established, and the XGBoost prediction model is trained using the training set data;

[0085] For the problem of multi-classification income level prediction, the same-level accuracy is taken as the optimization target to adjust the model parameters of the XGBoost prediction model.

[0086] In some embodiments, the processor 301 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 302, such as executing an intelligent prediction program for income capacity based on credit behavior characteristics.

[0087] This embodiment also provides a computer-readable storage medium, on which is stored a program for intelligently predicting income capacity based on credit behavior characteristics. When the program for intelligently predicting income capacity based on credit behavior characteristics is executed by a processor, the following steps are implemented:

[0088] Obtain customer credit performance, basic customer information and macro-indicator data, extract credit behavior characteristics, and establish a feature data set; divide the feature data set into a training set and a test set;

[0089] Taking income level as the prediction target, an XGBoost prediction model for income level prediction is established, and the XGBoost prediction model is trained using the training set data;

[0090] For the problem of multi-classification income level prediction, the same-level accuracy is taken as the optimization target to adjust the model parameters of the XGBoost prediction model.

[0091] The above-described embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the attached claims.

[0092] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent prediction method for income capacity based on credit behavior characteristics, characterized in that: include: Obtain customer credit performance, basic customer information and macro-indicator data, extract credit behavior characteristics, and establish a feature data set; Divide the feature dataset into training set and test set; Taking income level as the prediction target, an XGBoost prediction model for income level prediction is established, and the XGBoost prediction model is trained using the training set data; For the problem of multi-classification income level prediction, the same-level accuracy is taken as the optimization target to adjust the model parameters of the XGBoost prediction model.

2. The method for intelligently predicting income capacity based on credit behavior characteristics according to claim 1 is characterized in that: The method further comprises: Input the test set data into the trained XGBoost prediction model to obtain the income level prediction results; calculate the same level accuracy, upper and lower adjacent level accuracy, and upper and lower adjacent two levels accuracy based on the income level prediction results and actual income level to evaluate the accuracy of the model; Establish a model deployment and update process to ensure that the model can be put into production environment in a timely manner.

3. The method for intelligently predicting income capacity based on credit behavior characteristics according to claim 1 is characterized in that: After obtaining the customer's credit performance, basic customer information and macro-indicator data, the method further includes: Clean the data and handle missing values, outliers, and inconsistent data.

4. The method for intelligently predicting income capacity based on credit behavior characteristics according to claim 1 is characterized in that: The extracting of credit behavior characteristics includes: Credit behavior characteristics are extracted based on customer credit performance, basic customer information and macro-indicator data; among which, credit behavior characteristics include the customer's region, the nature of the customer's unit, occupational information and educational performance.

5. The method for intelligently predicting income capacity based on credit behavior characteristics according to claim 1 is characterized in that: Before establishing an XGBoost prediction model for income level prediction with income level as the prediction target, the method further includes: Assess the accuracy and coverage of different data sources in describing income levels and select the best data source; Income brackets were set based on sample averages.

6. The method for intelligently predicting income capacity based on credit behavior characteristics according to claim 1 is characterized in that: The above method takes the same gear accuracy as the optimization target and adjusts the model parameters of the XGBoost prediction model, specifically including: Through the Optuna parameter adjustment method, the hyperparameters of the XGBoost prediction model are tuned with the same-level accuracy as the optimization goal to obtain the optimal model parameter combination; among them, the hyperparameters of the XGBoost prediction model include tree depth, learning rate, and regularization coefficient.

7. An income capacity intelligent prediction system based on credit behavior characteristics applied to the income capacity intelligent prediction method based on credit behavior characteristics as described in any one of claims 1 to 6, characterized in that: include: Feature engineering module, which is used to obtain customer credit performance, basic customer information and macro-indicator data, extract credit behavior characteristics, and establish feature data sets; Divide the feature dataset into training set and test set; A model training module is used to establish an XGBoost prediction model for income level prediction with income level as the prediction target, and train the XGBoost prediction model using the training set data; The model optimization module is used to adjust the model parameters of the XGBoost prediction model for the multi-classification income level prediction problem, taking the same-level accuracy as the optimization target, and obtain the trained XGBoost prediction model.

8. An electronic device, It is characterized in that comprising a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the method for intelligently predicting income capacity based on credit behavior characteristics as described in any one of claims 1 to 6 above.

9. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the method for intelligent prediction of income capacity based on credit behavior characteristics as described in any one of claims 1 to 6 above.

Images