Credit card bill installment prediction method, device, storage medium and product

Abstract

The present application provides a credit card bill installment prediction method, device, storage medium and product. It is applied in the field of artificial intelligence technology. The specific implementation scheme is: obtaining the original data related to the bill installment corresponding to the target customer, the original data includes: customer basic data, customer behavior data, customer card account data and promotion feedback data; performing data preprocessing and feature expansion processing on the original data to obtain the target features related to the bill installment corresponding to the target customer; determining the bill installment intention score corresponding to the target customer based on the regression model trained to convergence and the target feature; if it is determined that the bill installment intention score is greater than the first preset score threshold, then determining the number of bill installments corresponding to the target customer based on the classification model trained to convergence and the target feature. The predicted bill installment intention score corresponding to the target customer is more accurate and can predict the number of bill installments corresponding to the target customer.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of artificial intelligence technology, and in particular to a credit card bill installment prediction method, device, storage medium and product. Background Art

[0002] As a payment tool that allows overdrafts, the number of credit cards issued has grown rapidly in recent years and has been widely recognized by consumers. In order to meet the consumption needs of consumers, credit cards can be used for bill installments. In order to better serve consumers, it is necessary to predict whether each consumer will use credit card bill installments, and notify consumers who are inclined to use bill installments of the news of bill installments to assist them in making bill installments.

[0003] The current method for predicting credit card bill installments is to obtain the consumer's basic information and historical bill information, extract features, and determine whether the consumer needs bill installments based on the extracted features.

[0004] However, the current method of predicting credit card bill installments does not fully consider the feedback from consumers when pushing bill installments, resulting in low accuracy in the prediction of credit card bill installments, and no prediction of the specific number of installments, resulting in the inability to provide consumers with more matching bill installment services. Summary of the invention

[0005] The embodiments of the present invention provide a credit card bill installment prediction method, device, storage medium and product to solve the technical problem that the prediction results of credit card bill installments in the prior art are low in accuracy and the specific installment number is not predicted, resulting in the inability to provide consumers with more matching bill installment services.

[0006] In a first aspect, an embodiment of the present invention provides a credit card bill installment prediction method, comprising:

[0007] Obtaining original data related to bill installment corresponding to target customers, the original data including: basic customer data, customer behavior data, customer card account data and promotion feedback data;

[0008] Performing data preprocessing and feature expansion processing on the original data to obtain target features related to bill installments corresponding to target customers;

[0009] Determine the bill installment intention score corresponding to the target customer based on the regression model trained to convergence and the target feature;

[0010] If it is determined that the bill installment intention score is greater than a first preset score threshold, the number of bill installments corresponding to the target customer is determined based on the classification model trained to convergence and the target feature.

[0011] In a second aspect, an embodiment of the present invention provides an electronic device, including: at least one processor and a memory;

[0012] The memory stores computer-executable instructions;

[0013] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method according to the first aspect.

[0014] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the method described in the first aspect.

[0015] In a fourth aspect, an embodiment of the present invention provides a computer program product, including a computer program, which implements the method described in the first aspect when executed by a processor.

[0016] The credit card bill installment prediction method, device, storage medium and product provided in the embodiments of the present invention obtain original data related to bill installments corresponding to target customers, wherein the original data includes: customer basic data, customer behavior data, customer card account data and promotion feedback data; the original data is subjected to data preprocessing and feature expansion processing to obtain target features related to bill installments corresponding to target customers; a bill installment intention score corresponding to the target customer is determined based on a regression model trained to convergence and the target features; if it is determined that the bill installment intention score is greater than a first preset score threshold, the number of bill installments corresponding to the target customer is determined based on a classification model trained to convergence and the target features. Since when predicting the credit card bill installment, not only the original data related to the bill installment, such as the customer basic data and the customer card account data, but also the original data related to the bill installment, such as the customer behavior data and the promotion feedback data, are obtained, after the original data is preprocessed and feature expanded, the target feature can more fully express the features related to the bill installment, and then after the target feature is regressed using the regression model trained to convergence, the predicted bill installment intention score corresponding to the target customer is more accurate, and thus the prediction result is more accurate. And after determining that the bill installment intention score is greater than the first preset score threshold, the classification model trained to convergence is used to classify the target feature, and then the number of bill installments corresponding to the target customer is determined, so that the number of bill installments corresponding to the target customer can be predicted, and then a more matching bill installment service can be provided to the target customer. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0018] Figure 1 This is a working principle diagram of the credit card bill installment prediction method that can implement the embodiment of the present invention;

[0019] Figure 2 It is a flow chart of a credit card bill installment prediction method provided by an embodiment of the present invention;

[0020] Figure 3 is a flow chart of a credit card bill installment prediction method provided by another embodiment of the present invention;

[0021] Figure 4 is a flow chart of a credit card bill installment prediction method provided by another embodiment of the present invention;

[0022] Figure 5 It is a flow chart of a credit card bill installment prediction method provided by another embodiment of the present invention;

[0023] Figure 6 is a schematic diagram of the structure of an electronic device used to implement the credit card bill installment prediction method of an embodiment of the present invention;

[0024] Figure 7 The block diagram is a block diagram of an electronic device used to implement the credit card bill installment prediction method of the embodiment of the present invention.

[0025] The above drawings show clear embodiments of the present disclosure, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present disclosure in any way, but to illustrate the concepts of the present disclosure to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0026] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0027] It should be noted that the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in the present invention are in compliance with the relevant laws and regulations and do not violate public order and good morals.

[0028] First, the terms involved in the embodiments of the present invention are explained:

[0029] Bill installment: The bill generated on the credit card billing date can be repaid in multiple installments. At the same time, the issuing bank will charge a certain installment fee.

[0030] XGBoost: The full name in English is: eXtreme Gradient Boosting, which is called extreme gradient boosting model in Chinese. It can handle various tasks such as regression, classification and sorting, supports parallel computing between nodes at the same level, has strong prediction performance and fast training speed.

[0031] LightGBM: A gradient boosting model based on the decision tree algorithm as the basic algorithm, usually used to implement classification tasks. LightGBM chooses to use Leaf-wise with depth restriction in leaf growth strategy, and the difference acceleration is implemented by histogram. It supports multi-threaded parallel processing, has better control over model complexity, and has a low probability of overfitting.

[0032] In order to clearly understand the technical solution of the present application, the solution of the prior art is introduced in detail.

[0033] The current method for predicting credit card bill installments is to extract features after obtaining the consumer's basic information and historical bill information, and then use a regression model or classification model to determine whether the consumer needs bill installments. Since the current method for predicting credit card bill installments only takes into account the consumer's basic information and historical bill information, and does not fully consider the consumer's feedback when pushing bill installments, it has not truly formed a closed loop in promotion, resulting in low accuracy in the prediction of credit card bill installments. Moreover, for consumers who have a strong intention to pay bills in installments, the number of bill installments for consumers is not predicted, resulting in the inability to provide consumers with more matching bill installment services.

[0034] Therefore, when facing the technical problems of the prior art, the inventor found through creative research that whether a customer makes bill installments is not only related to the customer's basic information and bill information, but also closely related to the customer's behavior data and promotion feedback data. Therefore, when obtaining the original data related to bill installments corresponding to the target customer, not only the customer's basic data and customer card account data are obtained, but also the customer's behavior data and promotion feedback data are obtained. After preprocessing and feature expansion processing of these data, the target features related to bill installments corresponding to the target customer are obtained, and the bill installment intention score corresponding to the target customer is determined according to the regression model trained to convergence and the target features. Since the target customer is very likely to make bill installments when the bill installment intention score is high, after determining that the bill installment intention score is greater than the first preset score threshold, the number of bill installments corresponding to the target customer is determined according to the classification model trained to convergence and the target feature. Since after determining the number of bill installments corresponding to the target customer, it is possible to configure matching activity plans for the target customer according to the different number of bill installments, it is possible to provide customers with more matching bill installment services.

[0035] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0036] Embodiment 1

[0037] Figure 2 FIG. 1 is a flow chart of a method for predicting the installment payment of a credit card bill provided by an embodiment of the present invention. Figure 2 As shown, the credit card bill installment prediction method provided in this embodiment is implemented by a credit card bill installment prediction device. The credit card bill installment prediction device is located in an electronic device. The credit card bill installment prediction method provided in this embodiment includes the following steps:

[0038] Step 201, obtaining original data related to bill installment corresponding to the target customer, the original data including: customer basic data, customer behavior data, customer card account data and promotion feedback data.

[0039] Among them, the basic customer data may include: customer age, customer gender, customer location, customer marital status, customer education level, customer ethnicity, etc.

[0040] Among them, customer behavior data refers to consumption behavior data on credit cards, such as transaction records, bill details, special installment records, etc.

[0041] Among them, customer card account data refers to credit card and account data, such as basic credit card data, basic credit card account data, basic special account data, etc. Among them, basic credit card data may include whether the credit card is expired, whether it is frozen, etc.

[0042] Among them, promotion feedback data refers to data related to customer feedback after bill installment promotion is carried out to customers, which may include: telephone promotion records, SMS promotion records, etc.

[0043] Optionally, in this embodiment, the original data related to the bill installment can be stored in the customer management platform. The electronic device sends a data acquisition request to the customer management platform, and the data acquisition request includes target customer identification information. The customer management platform obtains the original data related to the bill installment corresponding to the target customer based on the target customer identification information, and sends it to the electronic device, so that the electronic device obtains the original data related to the bill installment corresponding to the target customer.

[0044] The target customer identification information may be the target customer's account name, credit card number, or other information that represents a unique target customer.

[0045] Step 202, performing data preprocessing and feature expansion processing on the original data to obtain target features related to bill installments corresponding to the target customers.

[0046] In this embodiment, the original data may be pre-processed by data cleaning, missing value processing, format conversion, normalization, and continuous value discretization. The original data is then subjected to feature expansion processing according to a preset feature expansion strategy to obtain target features related to bill installments corresponding to the target customers.

[0047] Among them, the characteristics corresponding to the original data related to bill installments may include: basic customer characteristics, customer behavior characteristics, basic card account characteristics and promotion feedback characteristics.

[0048] Step 203, determining the bill installment intention score corresponding to the target customer based on the regression model trained to convergence and the target features.

[0049] The regression model trained to convergence may be an XGBoost model trained to convergence or other regression models trained to convergence, which is not limited in this embodiment.

[0050] In this embodiment, Figure 1 As shown, the target features are input into the regression model trained to convergence, and the target features are regressed using the regression model trained to convergence to determine the bill installment intention score corresponding to the target customer, and the bill installment intention score corresponding to the target customer is output.

[0051] Among them, the bill installment intention score corresponding to the target customer can be expressed as X.

[0052] It is understandable that the higher the bill installment intention score corresponding to the target customer, the greater the probability that the target customer will make bill installments. Conversely, the smaller the bill installment intention score corresponding to the target customer, the smaller the probability that the target customer will make bill installments.

[0053] Step 204: If it is determined that the bill installment intention score is greater than the first preset score threshold, the number of bill installments corresponding to the target customer is determined based on the classification model trained to convergence and the target features.

[0054] In this embodiment, the bill installment intention score corresponding to the target customer is compared with the first preset score threshold value X1. If it is determined that the bill installment intention score is greater than the first preset score threshold value (X>X1), the target customer is likely to make bill installments, and the corresponding bill installment period is determined for the target customer. Figure 1 As shown, the target features are input into the classification model trained to convergence, and the classification model trained to convergence is used to classify the target features to determine the number of bill installments corresponding to the target customers, and the number of bill installments corresponding to the target customers is output.

[0055] Among them, the classification model trained to convergence can be a LightGBM model trained to convergence or other classification models trained to convergence, which is not limited in this embodiment.

[0056] The credit card bill installment prediction method provided in this embodiment obtains the original data related to bill installment corresponding to the target customer, the original data includes: customer basic data, customer behavior data, customer card account data and promotion feedback data; performs data preprocessing and feature expansion processing on the original data to obtain the target features related to bill installment corresponding to the target customer; determines the bill installment intention score corresponding to the target customer according to the regression model trained to convergence and the target features; if it is determined that the bill installment intention score is greater than the first preset score threshold, determines the number of bill installments corresponding to the target customer according to the classification model trained to convergence and the target features. Since when predicting the credit card bill installment, not only the original data related to bill installment such as customer basic data and customer card account data are obtained, but also the original data related to bill installment such as customer behavior data and promotion feedback data are obtained, so after the original data is preprocessed and feature expanded, the target features can more fully express the features related to bill installment, and then after the target features are regressed using the regression model trained to convergence, the predicted bill installment intention score corresponding to the target customer is more accurate, and then the prediction result is more accurate. After determining that the bill installment intention score is greater than the first preset score threshold, the classification model trained to convergence is used to classify the target features, and then the number of bill installments corresponding to the target customer is determined. The number of bill installments corresponding to the target customer can be predicted, and then a more matching bill installment service can be provided to the target customer.

[0057] Embodiment 2

[0058] Figure 3 FIG. 2 is a flow chart of a credit card bill installment prediction method provided by another embodiment of the present invention. Figure 3 As shown, the credit card bill installment prediction method provided in this embodiment is based on the credit card bill installment prediction method provided in Example 1, and further refines steps 202 to 204, and also includes other steps. The credit card bill installment prediction method provided in this embodiment includes the following steps:

[0059] Step 301, obtaining original data related to bill installment corresponding to the target customer, the original data including: customer basic data, customer behavior data, customer card account data and promotion feedback data.

[0060] Step 302, performing missing value processing on the original data; obtaining first preset data in the original data; and performing one-hot encoding processing on the first preset data.

[0061] In this embodiment, step 302 is an optional implementation method of performing data preprocessing on the original data in step 202 of the first embodiment.

[0062] Specifically, in this embodiment, in order to facilitate the next step of processing of the original data, the missing or empty original data is uniformly processed with missing values, such as replacing all missing or empty original data with missing values, and the missing values ​​may be NA.

[0063] In this embodiment, discrete data in the original data is obtained as the first preset data, and the first preset data is normalized. When performing the normalization process, the first preset data is one-hot encoded using a one-hot encoding algorithm (English: One-HotEncoding).

[0064] The first preset data may include the customer's gender, the customer's location, the customer's marital status, the customer's education level, the customer's ethnicity, etc. in the customer's basic data. The first preset data may also include the repayment method in the customer's card account data.

[0065] Since the basic customer data does not need to be expanded after data preprocessing, the target features corresponding to the basic customer data of the target customers may include: [age, gender one-hot coding, region one-hot coding, marital status one-hot coding, education level one-hot coding, ethnicity one-hot coding]

[0066] Step 303, obtaining second preset data and a corresponding feature expansion strategy in the original data; performing feature expansion processing on the second preset data according to the corresponding feature expansion strategy.

[0067] In this embodiment, step 303 is an optional implementation method of performing feature expansion processing on the original data in step 202 of the first embodiment.

[0068] In this embodiment, the feature expansion processing is to obtain feature values ​​such as variance, mode, maximum value, minimum value, or coefficient of variation from the data according to the corresponding feature expansion strategy.

[0069] In this embodiment, in the customer behavior data of the original data, the second preset data includes the current period consumption data and the historical installment data in the bill details. After the second preset data is subjected to feature expansion processing, the number of current period consumption and the mode of the historical installment periods are obtained. The bill amount, the current period handling fee amount, the historical installment principal, and the historical installment repayment amount for each period are obtained from the customer behavior data, and then the target feature corresponding to the customer behavior data is determined. The target feature corresponding to the customer behavior data is the target customer behavior feature, and the target customer behavior feature may include: [bill amount, current period handling fee amount, number of current period consumption, mode of the historical installment periods, historical installment principal, and historical installment repayment amount for each period]. Among them,

[0070] In this embodiment, in the customer card account data, the second preset data includes: the account billing situation and the consumption balance of each period in the basic data of the credit card account. After the second preset data is subjected to feature expansion processing according to the corresponding feature expansion strategy, the number of consecutive periods in which the account has not been billed is obtained according to the account billing situation, and the current additional consumption balance is determined according to the consumption balance of each period. The maximum card limit, the maximum account limit, the repayment method code, the overdraft mark of this billing cycle and other data are obtained from the customer card account data, and then the target feature corresponding to the customer card account data is determined. The target feature corresponding to the customer card account data is the target customer card account feature, and the target customer card account feature may include: [maximum card limit, maximum account limit, number of consecutive periods in which the account has not been billed, repayment method code, overdraft mark of this billing cycle, and current additional consumption balance].

[0071] In this embodiment, in the promotion feedback data, the second preset data includes: telephone promotion records and SMS promotion records. After data preprocessing is performed on the promotion feedback data and the second preset data is subjected to feature expansion processing according to the corresponding feature expansion strategy, the target feature corresponding to the promotion feedback data of the target customer is the target promotion feedback feature, and the target promotion feedback feature may include: [number of phone calls, average call duration, longest call duration, number of days from the last call, number of SMS marketing, number of SMS unsubscriptions, number of marketing successes, and marketing success rate].

[0072] Step 304, input the target feature data into the regression model trained to convergence; use the regression model trained to convergence to perform regression processing on the target feature to determine and output the bill installment intention score corresponding to the target customer.

[0073] Optionally, in this embodiment, the classification model trained to convergence is an XGBoost model trained to convergence.

[0074] In this embodiment, the target feature data is input into the XGBoost model trained to convergence, and the XGBoost model trained to convergence performs regression processing on the target feature to determine the bill installment intention score corresponding to the target customer, and outputs the bill installment intention score corresponding to the target customer. Since the XGBoost model can support parallel computing between nodes at the same level, the prediction speed of the bill installment intention score corresponding to the target customer is accelerated.

[0075] Step 305: Perform corresponding bill installment promotion according to the bill installment intention score corresponding to the target customer.

[0076] Optionally, step 305 includes the following solutions:

[0077] If it is determined that the bill installment intention score is less than the second preset score threshold, the target customer identification information is sent to the application information prompt device, so that the application information prompt device sends an information prompt message to the target customer client.

[0078] If it is determined that the bill installment intention score is greater than or equal to the third preset score threshold and less than the fourth preset score threshold, the target customer identification information is sent to the SMS prompt device, so that the SMS prompt device sends an SMS prompt message to the target customer terminal.

[0079] If it is determined that the bill installment intention score is greater than or equal to the fourth preset score threshold, the target customer identification information is sent to the outbound call device, so that the outbound call device makes a call to the target customer terminal.

[0080] Among them, the second preset score threshold is represented as X2, the third preset score threshold is represented as X3, and the fourth preset score threshold is represented as X4. Compare the bill installment intention score X corresponding to the target customer with each preset score threshold. If X < X2, it means that the probability of the target customer applying for bill installment is slightly higher, and the target customer identification information is sent to the application information prompt device, so that the application information prompt device sends an information prompt message to the target customer client. If X2 ≤ X < X3, it means that the probability of the target customer applying for bill installment is relatively high, and the target customer identification information is sent to the SMS prompt device, so that the SMS prompt device sends an SMS prompt message to the target customer terminal. If X3 ≤ X ≤ X4, it means that the probability of the target customer applying for bill installment is very high. The target customer identification information is sent to the outbound call device, so that the outbound call device makes a call to the target customer terminal to promote bill installment for the target user in a targeted manner. Since personalized bill installment promotion plans are provided for users with bill installment intention scores at different stages, the success rate of target customers applying for bill installment can be effectively improved.

[0081] Step 306, if it is determined that the bill installment intention score is greater than the first preset score threshold, input the target features into the classification model trained to convergence; use the classification model trained to convergence to classify the target features to determine and output the number of installments of the bill installment corresponding to the target customer.

[0082] Optionally, in this embodiment, the classification model trained to convergence is a LightGBM model trained to convergence. Among them, each category corresponds to a number of installments of the bill installment.

[0083] In this embodiment, if it is determined that the bill installment intention score is greater than the first preset score threshold, it means that the target customer is very likely to apply for bill installment, and the corresponding number of installments of the bill installment is determined for the target customer.

[0084] Specifically, when determining the corresponding bill installment number for the target customer, the target feature is input into the LightGBM model trained to convergence, and the LightGBM model trained to convergence is used to classify the target feature, determine the bill installment number corresponding to the target customer, and output the bill installment number corresponding to the target customer. Since the LightGBM model supports multi-threaded parallel processing, the speed of determining the corresponding bill installment number for the target customer is accelerated.

[0085] Embodiment 3

[0086] Figure 4 FIG. 1 is a flow chart of a credit card bill installment prediction method provided by another embodiment of the present invention. Figure 4 As shown, the credit card bill installment prediction method provided in this embodiment is based on the credit card bill installment prediction method provided in any of the above embodiments, and further includes the steps of training a preset regression model and a preset classification model. The credit card bill installment prediction method provided in this embodiment also includes the following steps:

[0087] Step 401: Obtain a first training sample set for training a preset regression model and a second training sample set for training a preset classification model.

[0088] Among them, the second training sample in the second training sample set is the first training sample in the first training sample set whose bill installment intention score is greater than the first preset score threshold.

[0089] In this embodiment, the first training sample set is a sample feature set corresponding to each first sample customer. That is, each first training sample is a sample feature corresponding to each first sample customer. Similarly, the second training sample set is a sample feature set corresponding to each second sample customer. That is, each second training sample is a sample feature corresponding to each second sample customer.

[0090] Among them, the second sample customer is the first sample customer whose bill installment intention score is greater than the first preset score threshold.

[0091] Specifically, the sample characteristics corresponding to the first sample customers include: basic customer characteristics, basic card account characteristics, customer behavior characteristics, and promotion feedback characteristics. Similarly, the sample characteristics corresponding to the second sample customers also include: basic customer characteristics, basic card account characteristics, customer behavior characteristics, and promotion feedback characteristics.

[0092] It should be noted that the implementation method of determining the sample characteristics corresponding to the first sample customers and determining the sample characteristics corresponding to the second sample customers is similar to the implementation method of determining the target characteristics related to bill installment corresponding to the target customers in the above embodiment, and will not be repeated here.

[0093] Step 402: Use the first training sample to train the preset regression model to obtain a regression model that is trained to convergence.

[0094] Among them, the preset regression model is the preset XGBoost model.

[0095] In this embodiment, the sample features corresponding to the first sample customer are input into the preset XGBoost model, and the preset XGBoost model is used to perform regression processing on the sample features corresponding to the first sample customer to train the preset XGBoost model. During the training process, it is determined whether the preset training convergence condition is met. If it is determined that the preset convergence condition is not met, the training parameters are adjusted and the XGBoost model is continued to be trained until the preset convergence condition is met.

[0096] The preset convergence condition may be that the loss function reaches a minimum, or reaches a preset convergence number, etc., which is not limited in this embodiment.

[0097] Step 403: Use the second training sample to train the preset classification model to obtain a classification model that is trained to convergence.

[0098] Among them, the preset classification model is the preset LightGBM model.

[0099] In this embodiment, the sample features corresponding to the second sample customers are input into the preset LightGBM model, and the preset LightGBM model is used to classify the sample features corresponding to the second sample customers to train the preset LightGBM model. During the training process, it is determined whether the preset training convergence condition is met. If it is determined that the preset convergence condition is not met, the training parameters are adjusted and the LightGBM model is continued to be trained until the preset convergence condition is met.

[0100] The preset convergence condition may be that the loss function reaches a minimum, or reaches a preset convergence number, etc., which is not limited in this embodiment.

[0101] The credit card bill installment prediction method provided in this embodiment can speed up the training of the regression model when the preset regression model is trained using the first training sample to obtain a regression model that is trained to convergence, because the XGBoost model supports parallel computing between nodes at the same level.

[0102] The credit card bill installment prediction method provided in this embodiment has better control over the complexity of the LightGBM model and a lower probability of overfitting. Therefore, when the second training sample is used to train the preset classification model to obtain a classification model that is trained to convergence, a LightGBM model with better performance can be obtained.

[0103] Embodiment 4

[0104] Figure 5 FIG. 1 is a flow chart of a credit card bill installment prediction method provided by another embodiment of the present invention. Figure 5 As shown, the credit card bill installment prediction method provided in this embodiment is based on the credit card bill installment prediction method provided in the fourth embodiment, and after determining the number of bill installments corresponding to the target customer according to the classification model trained to convergence and the target features, it also includes the following steps:

[0105] Step 501, obtaining activity plan data for multiple bill installment configurations.

[0106] In this embodiment, after determining the number of bill installments corresponding to a plurality of target customers, the professional may configure activity plan data for each number of bill installments.

[0107] For example, if the number of bill installments is 6 installments, 12 installments and 24 installments, then the activity plan data of 6 installments, 12 installments and 24 installments are configured specifically.

[0108] The activity plan data for each bill installment configuration may include characteristic values ​​such as a handling fee rate and a handling fee amount.

[0109] Specifically, in this embodiment, an activity plan data configuration page can be provided to professionals. Through this configuration page, professionals configure corresponding activity plan data for the number of installments of various bills, and then the electronic device obtains the activity plan data configured for the number of installments of various bills.

[0110] Step 502, determining the bill installment success rates of target customers corresponding to the multiple bill installment numbers.

[0111] In this embodiment, the results of whether each target customer performs bill installment payment corresponding to the multiple bill installment payment periods are obtained, and the bill installment success rate of the target customer is determined according to the results of whether the bill installment payment is performed.

[0112] Step 503: If it is determined that the bill installment success rate is less than the preset success rate threshold, the activity plan data is adjusted, and the adjusted activity plan data is added or updated to the first training sample and the second training sample.

[0113] In this embodiment, the bill installment success rate is compared with the preset success rate threshold. If it is determined that the bill installment success rate is greater than the preset success rate threshold, it means that the corresponding customer is relatively acceptable to the activity plan data for the multiple bill installment period configurations, so the bill installment is performed. If it is determined that the bill installment success rate is less than or equal to the preset success rate threshold, it means that the corresponding customer is not very acceptable to the activity plan data for the multiple bill installment period configurations, so the activity plan data needs to be adjusted.

[0114] Among them, when adjusting the activity plan data, professionals can appropriately reduce characteristic values ​​such as the handling fee rate and handling fee amount for the activity plan data corresponding to the number of installments for each bill in the activity plan data configuration page.

[0115] In this embodiment, since the activity plan data is also data that is highly correlated with the bill installment, if the activity plan data cannot be obtained when the preset regression model and the preset classification model are trained, the adjusted activity plan data is added to the first training sample and the second training sample. If the activity plan data can be obtained when the preset regression model and the preset classification model are trained, the original activity plan data in the first training sample and the second training sample is updated.

[0116] Step 504: optimize the regression model trained to convergence according to the first training sample, and optimize the classification model trained to convergence according to the second training sample.

[0117] In this embodiment, the first training sample including the added or updated adjusted activity plan data is used to continue to tune the regression model trained to convergence, so that the regression model trained to convergence is adapted to the adjusted activity plan data. Similarly, the second training sample including the added or updated adjusted activity plan data is used to continue to tune the classification model trained to convergence, so that the classification model trained to convergence is adapted to the adjusted activity plan data.

[0118] The credit card bill installment prediction method provided in this embodiment obtains activity plan data configured for multiple bill installment periods; determines the bill installment success rate of target customers corresponding to the multiple bill installment periods; if it is determined that the bill installment success rate is less than a preset success rate threshold, adjusts the activity plan data, and adds or updates the adjusted activity plan data to the first training sample and the second training sample; optimizes the regression model trained to convergence according to the first training sample, and optimizes the classification model trained to convergence according to the second training sample, and when using the optimized regression model trained to convergence to predict whether the target customer will make a bill installment, it can predict that more target customers will make a bill installment, and then after corresponding promotion, more target users can make a bill installment, thereby promoting the continuous development of the bill installment business.

[0119] It is understandable that the original data related to bill installments corresponding to the target customers may also include corresponding activity plan data, and preprocessing and feature expansion of the activity plan data can enable the target features to more fully express the characteristics of the target users related to bill installments, thereby further improving the accuracy of the prediction.

[0120] Embodiment 5

[0121] Figure 6 Schematic diagram of the structure of an electronic device for implementing the credit card bill installment prediction method according to an embodiment of the present invention. Figure 6 As shown, the electronic device 60 provided in this embodiment includes: at least one processor 62 and a memory 61.

[0122] The memory 61 stores computer-executable instructions. At least one processor 62 executes the computer-executable instructions stored in the memory, so that at least one processor executes the method provided in any one of the first to fourth embodiments.

[0123] Embodiment 6

[0124] Figure 7 is a block diagram of an electronic device for implementing the credit card bill installment prediction method according to an embodiment of the present invention, such as Figure 7 As shown, the electronic device can be a computer, a messaging device, a tablet device, a server, a server cluster, etc.

[0125] The electronic device 700 may include one or more of the following components: a processing component 702 , a memory 704 , a power component 706 , an input / output (I / O) interface 712 , a sensor component 714 , and a communication component 716 .

[0126] The processing component 702 generally controls the overall operation of the electronic device 700, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 702 may include one or more processors 720 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 702 may include one or more modules to facilitate the interaction between the processing component 702 and other components.

[0127] The memory 704 is configured to store various types of data to support operations on the electronic device 700. Examples of such data include instructions for any application or method operating on the electronic device 700, contact data, phone book data, messages, pictures, videos, etc. The memory 704 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0128] The power supply component 706 provides power to the various components of the electronic device 700. The power supply component 706 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the electronic device 700.

[0129] I / O interface 712 provides an interface between processing component 702 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: home button, volume button, start button, and lock button.

[0130] The sensor assembly 714 includes one or more sensors for providing various aspects of status assessment for the electronic device 700. For example, the sensor assembly 714 can detect the open / closed state of the electronic device 700, the relative positioning of components, such as the display and keypad of the electronic device 700, and the sensor assembly 714 can also detect the position change of the electronic device 700 or a component of the electronic device 700, the presence or absence of user contact with the electronic device 700, the orientation or acceleration / deceleration of the electronic device 700, and the temperature change of the electronic device 700. The sensor assembly 714 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 714 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 714 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0131] The communication component 716 is configured to facilitate wired or wireless communication between the electronic device 700 and other devices. The electronic device 700 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 716 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 716 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0132] In an exemplary embodiment, the electronic device 700 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above methods.

[0133] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 704 including instructions, and the above instructions can be executed by the processor 720 of the electronic device 700 to complete the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0134] A non-temporary computer-readable storage medium, when the instructions in the storage medium are executed by a processor of an electronic device, enables the electronic device to execute the method provided by any one of the above-mentioned embodiments 1 to 3.

[0135] In an exemplary embodiment, a computer program product is also provided, including a computer program, where a processor executes the method provided by any one of the above-mentioned embodiments 1 to 3.

[0136] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be regarded as exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0137] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A credit card bill installment prediction method, characterized in that: include: Obtaining original data related to bill installment corresponding to target customers, the original data including: basic customer data, customer behavior data, customer card account data and promotion feedback data; the basic customer data includes: customer age, customer gender, customer location, customer marital status, customer education level, customer ethnicity; the customer behavior data refers to credit card consumption behavior data; the customer card account data refers to credit card and account data; the promotion feedback data refers to data related to customer feedback after bill installment promotion is carried out on customers; Performing data preprocessing and feature expansion processing on the original data to obtain target features related to bill installments corresponding to target customers; Determine the bill installment intention score corresponding to the target customer based on the regression model trained to convergence and the target feature; If it is determined that the bill installment intention score is greater than a first preset score threshold, determining the number of bill installments corresponding to the target customer based on the classification model trained to convergence and the target feature; After determining the bill installment intention score corresponding to the target customer according to the regression model trained to convergence and the target feature, the method further includes: If it is determined that the bill installment intention score is less than a second preset score threshold, sending the target customer identification information to an application information prompting device, so that the application information prompting device sends an information prompting message to the target customer client; If it is determined that the bill installment intention score is greater than or equal to the second preset score threshold and less than the third preset score threshold, sending the target customer identification information to a text message prompt device so that the text message prompt device sends a text message prompt information to the target customer terminal; If it is determined that the bill installment intention score is greater than or equal to the third preset score threshold and less than or equal to the fourth preset score threshold, the target customer identification information is sent to the outbound call device so that the outbound call device calls the target customer terminal.

2. The method according to claim 1, characterized in that The data preprocessing of the original data comprises: Performing missing value processing on the original data; Acquire first preset data from the original data; the first preset data includes the customer's gender, the customer's region, the customer's marital status, the customer's education level, and the customer's ethnicity in the customer's basic data; The first preset data is subjected to one-hot encoding processing.

3. The method according to claim 1 or 2, characterized in that: Performing feature expansion processing on the original data includes: Obtaining second preset data and corresponding feature expansion strategies from the original data; the second preset data includes current period consumption data, historical installment data, account billing information, each period consumption balance, telephone promotion records, and SMS promotion records in the bill details; The second preset data is subjected to feature expansion processing according to a corresponding feature expansion strategy.

4. The method according to claim 1, characterized in that: The regression model trained to convergence is an XGBoost model trained to convergence; The step of determining the bill installment intention score corresponding to the target customer based on the regression model trained to convergence and the target feature includes: Inputting the target feature data into a regression model trained to convergence; The target feature is regressed using the regression model trained to convergence to determine and output the bill installment intention score corresponding to the target customer.

5. The method according to claim 1, characterized in that The classification model trained to convergence is a LightGBM model trained to convergence; The step of determining the number of bill installments corresponding to the target customer based on the classification model trained to convergence and the target feature includes: Inputting the target features into a classification model trained to convergence; The target features are classified using the classification model trained to convergence to determine and output the number of bill installments corresponding to the target customer.

6. The method according to claim 1, characterized in that Before determining the bill installment intention score corresponding to the target customer according to the regression model trained to convergence and the target feature, the method further includes: Obtaining a first training sample set for training a preset regression model and a second training sample set for training a preset classification model, wherein the second training sample in the second training sample set is the first training sample in the first training sample set whose bill installment intention score is greater than a first preset score threshold; Using the first training sample to train the preset regression model to obtain a regression model that is trained to convergence; The preset classification model is trained using the second training sample to obtain a classification model that is trained to convergence.

7. The method according to claim 6, characterized in that After determining the number of bill installments corresponding to the target customer according to the classification model trained to convergence and the target feature, the method further includes: Obtain activity plan data for various bill installment configurations; Determine the bill installment success rate of target customers corresponding to the number of bill installments; If it is determined that the bill installment success rate is less than a preset success rate threshold, adjusting the activity plan data, and adding or updating the adjusted activity plan data to the first training sample and the second training sample; The regression model trained to convergence is tuned according to the first training sample, and the classification model trained to convergence is tuned according to the second training sample.

8. An electronic device, characterized in that: include: at least one processor and memory; The memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images