Method and device for processing capital loss of client issuing on behalf, and electronic equipment

Abstract

The invention discloses a method and a device for processing capital loss of a consigned customer, and electronic equipment. The method comprises the following steps: acquiring basic information data of a target customer and fund issuing data in a preset period; inputting the basic information data and the fund issuing data into a fund loss prediction model to obtain a predicted fund loss rate of the target customer on the target date; under the condition that the predicted fund loss rate is greater than or equal to a fund loss rate threshold value, determining the target customer as a fund loss customer; and determining a target fund saving strategy based on the basic information data and the fund issuing data, and executing the target fund saving strategy for the fund loss customer. Through the method and the device, the problem of relatively high capital loss of the consigned customer caused by inaccurate prediction of capital loss of the consigned customer in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and more specifically, to a method, device and electronic device for processing fund loss of a client who is sending funds on behalf of another. Background Art

[0002] The clientele of financial institutions who receive payroll payments have problems such as low retention rate of funds and low contribution from payroll customers. This is because the frontline branches of financial institutions cannot fully grasp the payroll time and are unable to coordinate customer maintenance work. In addition, the time when customers' payroll funds are lost is unclear, and the branches do not understand customers' fund usage habits and are unable to provide timely on-site maintenance.

[0003] In the related art, the method of dealing with the loss of funds of the agency customers is to rely on manual analysis of customer transaction behavior and screening of the agency customers, which is not only inefficient but also difficult to guarantee accuracy. The use of machine learning methods in predicting the loss of funds of agency customers also has limitations. The fund loss prediction methods in the related art are not accurate enough. When the outlets try to retain the funds according to the loss list, some of the agency funds have already been transferred away.

[0004] Currently, no effective solution has been proposed to solve the problem that the prediction of fund loss of agency payee customers in related technologies is not accurate enough, resulting in high fund loss of agency payee customers. Summary of the invention

[0005] The main purpose of this application is to provide a method, device and electronic equipment for processing the loss of funds of an agent-based customer, so as to solve the problem that the prediction of the loss of funds of an agent-based customer in the related technology is not accurate enough, resulting in a high loss of funds of the agent-based customer.

[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for processing the loss of funds of a client who obtains funds on behalf of another party is provided. The method comprises: obtaining basic information data of a target client and data on funds on behalf of another party within a preset period, wherein the basic information data comprises at least one of the following: identity information data, funds distribution data and user behavior data, and the data on funds on behalf of another party comprises at least one of the following: average amount of funds on behalf of another party, maximum amount of funds on behalf of another party and average amount of loss; inputting the basic information data and the data on funds on behalf of another party into a fund loss prediction model to obtain the predicted fund loss rate of the target client on the target date, wherein the fund loss prediction model is obtained by training with multiple groups of training samples, each group of training samples comprises basic information data, historical fund loss data and fund loss rate, and the target date is the date after the preset number of days when the target client obtains funds on behalf of another party; in the case where the predicted fund loss rate is greater than or equal to the fund loss rate threshold, the target client is determined as a client who has lost funds; based on the basic information data and the data on funds on behalf of another party, a target fund retention strategy is determined, and the target fund retention strategy is executed on the client who has lost funds.

[0007] Optionally, the funds loss prediction model is trained in the following manner: determine the group of agency customers, obtain basic information data and historical funds payment data of the agency customer group; determine the funds loss rate of each agency customer on the target date from the historical funds payment data; determine the basic information data, historical funds payment data and funds loss rate of each agency customer as a group of training samples, and obtain multiple groups of training samples for the agency customer group; train the machine learning model based on the multiple groups of training samples to obtain the funds loss prediction model.

[0008] Optionally, determining a target fund retention strategy based on basic information data and fund payment data includes: determining similar customer groups of target customers based on basic information data and fund payment data; obtaining multiple fund retention strategies executed by similar customer groups within a target period, and determining the retention success rate of each fund retention strategy to obtain a set of retention success rates; and determining the fund retention strategy corresponding to the maximum retention success rate in a set of retention success rates as the target fund retention strategy.

[0009] Optionally, determining similar customer groups of target customers based on basic information data and fund payment data includes: obtaining basic information data and fund payment data of each fund payment customer, converting the basic information data and fund payment data of each fund payment customer into a first feature vector to obtain a group of first feature vectors; converting the basic information data and fund payment data of the target customer into a second feature vector; calculating the similarity between the second feature vector and each first feature vector; determining the fund payment customer corresponding to the first feature vector whose similarity is greater than or equal to a similarity threshold as a similar customer, and defining the set of all identified similar customers as a similar customer group.

[0010] Optionally, after executing the target fund retention strategy for the fund-losing customers, the method also includes: determining the actual fund loss rate of the target customers on the target date, and judging whether the actual fund loss rate is greater than or equal to a fund loss rate threshold; when the actual fund loss rate is greater than or equal to the fund loss rate threshold, determining that the target fund retention strategy has failed to retain the customers; when the actual fund loss rate is less than the fund loss rate threshold, determining that the target fund retention strategy has succeeded to retain the customers; and updating the retention success rate of the target fund retention strategy based on the retention result of the target fund retention strategy.

[0011] Optionally, after obtaining the predicted capital churn rate of the target customer on the target date, the method also includes: when the predicted capital churn rate is less than a capital churn rate threshold, determining the actual capital churn rate of the target customer on the target date, and judging whether the actual capital churn rate is greater than or equal to the capital churn rate threshold; when the actual capital churn rate is less than the capital churn rate threshold, determining that the capital churn rate prediction result of the target customer is accurate; when the actual capital churn rate is greater than or equal to the capital churn rate threshold, determining that the capital churn rate prediction result of the target customer is wrong; and calculating the accuracy of the capital churn rate prediction result of the capital churn prediction model within the statistical model update cycle, and updating the capital churn prediction model when the accuracy is less than the accuracy threshold.

[0012] Optionally, before determining that the predicted result of the target customer's capital churn rate is accurate, the method also includes: calculating the difference between the predicted capital churn rate and the actual capital churn rate, and determining whether the difference is greater than or equal to a preset numerical threshold; when the difference is greater than or equal to the preset numerical threshold, determining that the predicted result of the target customer's capital churn rate is wrong; when the difference is less than the preset numerical threshold, determining that the predicted result of the target customer's capital churn rate is accurate.

[0013] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a device for processing the loss of funds of a client who has received funds on behalf of another party is provided. The device comprises: an acquisition unit, which is used to acquire the basic information data of the target client and the fund payment data within a preset period, wherein the basic information data includes at least one of the following: identity information data, fund distribution data and user behavior data, and the fund payment data includes at least one of the following: average payment amount, maximum payment amount and average loss amount; an input unit, which is used to input the basic information data and the fund payment data into a fund loss prediction model to obtain the predicted fund loss rate of the target client on the target date, wherein the fund loss prediction model is obtained by training multiple groups of training samples, each group of training samples includes basic information data, historical fund payment data and fund loss rate, and the target date is the date after the preset number of days when the target client obtains the funds on behalf of another party; a first determination unit, which is used to determine the target client as a client who has lost funds when the predicted fund loss rate is greater than or equal to the fund loss rate threshold; an execution unit, which is used to determine the target fund retention strategy based on the basic information data and the fund payment data, and execute the target fund retention strategy on the client who has lost funds.

[0014] In an embodiment of the present application, basic information data of a target customer and fund payment data within a preset period are obtained, wherein the basic information data includes at least one of the following: identity information data, fund distribution data and user behavior data, and the fund payment data includes at least one of the following: average payment amount, maximum payment amount and average loss amount; the basic information data and fund payment data are input into a fund loss prediction model to obtain a predicted fund loss rate of the target customer on a target date, wherein the fund loss prediction model is trained by multiple groups of training samples, each group of training samples includes basic information data, historical fund payment data and fund loss rate, and the target date is a date after a preset number of days from when the target customer obtains the payment funds; in the predicted fund loss rate When the capital churn rate is greater than or equal to the capital churn rate threshold, the target customer is determined as a capital churn customer; the target capital retention strategy is determined based on the basic information data and the fund payment data, and the target capital retention strategy is executed on the capital churn customer. The customer's capital churn rate on the target date is predicted through the customer's basic information data and the fund payment data, and then the customer is evaluated whether he is a capital churn customer based on the capital churn rate, and then the target capital retention strategy is executed on the capital churn customer, thereby achieving the purpose of improving the accuracy of predicting capital churn for fund payment customers, thereby achieving the technical effect of improving the success rate of recovering capital churn for fund payment customers, and then solving the technical problem that the prediction of capital churn for fund payment customers is not accurate enough, resulting in a high capital churn for fund payment customers. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0016] Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for processing the loss of funds of a proxy payment client is shown;

[0017] Figure 2 It is a flow chart of a method for processing fund loss of a client according to an embodiment of the present application;

[0018] Figure 3 It is a flowchart of an optional method for processing loss of funds of a client for agency payment provided in an embodiment of the present application;

[0019] Figure 4 It is a schematic diagram of the execution process of the client funds retention strategy recommendation algorithm provided according to the embodiment of the present application;

[0020] Figure 5 It is a schematic diagram of a device for processing fund loss of a client provided on behalf of another according to an embodiment of the present application;

[0021] Figure 6 It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] It should be noted that the collected information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this application are information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with relevant laws, regulations and standards, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation entrances are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or institutions to provide users with corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process will be entered.

[0025] Example 1

[0026] According to an embodiment of the present application, a method embodiment for handling the loss of funds of a client who is to be paid on behalf of another is also provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0027] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for processing the loss of funds of a client who is paying on behalf of another party is shown. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more (shown as 102a, 102b, ..., 102n in the figure) processors 102 (the processor 102 may include but is not limited to a processing device such as an MCU (Microcontroller Unit) or an FPGA (Field-Programmable Gate Array), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a USB (Universal Serial Bus) port (which may be included as one of the ports of a BUS (Business)), a network interface, a power supply and / or a camera. It will be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0028] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0029] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the method for processing the loss of funds of the agency-issued customer in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the above-mentioned method for processing the loss of funds of the agency-issued customer is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0030] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0031] The display may be, for example, a touch screen liquid crystal display that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0032] In the above operating environment, this application provides a method for handling the loss of funds of a client who is paying on behalf of another. Figure 2 : is a flow chart of a method for processing fund loss of a client according to an embodiment of the present application, such as Figure 2 As shown, the method includes:

[0033] Step S201, obtaining basic information data of the target customer and fund payment data within a preset period, wherein the basic information data includes at least one of the following: identity information data, fund distribution data and user behavior data, and the fund payment data includes at least one of the following: average payment amount, maximum payment amount and average loss amount.

[0034] In step S201, the identity information data may include data such as age, gender, occupation, education level, marital status, and place of residence. The fund distribution data may include deposit balance, loan information, investment status, etc. The user behavior data may include transaction frequency, transaction amount, transaction type, account usage habits, customer feedback, etc. The average amount of payment may be the average amount of payment for the target customer within a preset period (such as the past n months). The maximum amount of payment is the maximum amount of payment within the preset period. The average amount of loss is the average amount of loss within the preset period.

[0035] After obtaining these data, preprocessing steps such as data cleaning, deduplication, outlier detection and missing value filling are performed to ensure data quality. Then, feature engineering methods are used to transform and derive the data, for example, deriving age level features from age data and calculating capital loss rate from transaction amount. These processed data will be used as input for model training and prediction of capital loss rate.

[0036] For example, based on the experience of the agency payment business experts, we can explore the factors that affect the loss of assets of the agency payment customer group and design a feature wide table containing data feature variables, mainly covering the following aspects: basic information, asset distribution, behavioral habits, product stickiness, equity preference and other k features. In addition, we can construct statistical features, such as the average monthly loss amount of agency payment funds in the past n months, Miss ave , the maximum amount of funds sent on behalf of others in the past n months MAX acc The average monthly loss of funds on behalf of others in the past n months is Miss ave Calculated by the following formula:

[0037]

[0038] Among them, account i is the amount of payment made by the customer in the ith month, missing i is the amount of customer loss in the ith month. The average monthly amount of customer payment in the past n months is A ave It can be calculated by the following formula:

[0039]

[0040] Among them, the maximum amount of funds sent on behalf of others in the past n months is MAX acc It can be calculated by the following formula:

[0041] MAX acc =max(account i , i∈{1,2,3...,n});

[0042] Obtain the historical data of the customer and other relevant feature data to form the feature set T k, the final dataset wide table is as follows: T i = {P i ,account i ,Send_date i ,T1 i ...Tk i ,Miss ave ,A ave ,MAX acc}, where P i Represents the customer number, account of the i-th month i Represents the customer's payment amount in month i, Send_date i Represents the date of the i-th month. Obtain the historical data of the client within the set time window (that is, the preset period), including basic information, asset distribution, behavioral habits, product stickiness and other features. Perform data cleaning, data processing and feature screening on the historical data. Various indicator features form a data wide table. Perform data preprocessing operations on the data set, remove duplicate values ​​and abnormal value data samples in the data wide table, and fill in missing values. At the same time, perform feature vectorization processing on features such as age. Using the random forest model, the importance of the feature is evaluated by calculating the contribution of each feature on each tree. Using the Gini index as an evaluation indicator, the features that have the greatest impact on the model prediction results are identified in the data set with a large number of features, forming the input feature set T of the client fund loss prediction model. Through methods such as the chi-square test, the correlation between variables is eliminated, and finally the feature set related to the prediction of client fund loss is screened out.

[0043] Step S202, input the basic information data and the fund payment data into the fund loss prediction model to obtain the predicted fund loss rate of the target customer on the target date, wherein the fund loss prediction model is trained by multiple groups of training samples, each group of training samples includes basic information data, historical fund payment data and fund loss rate, and the target date is the date after the preset number of days for the target customer to obtain the payment funds.

[0044] In step S202, a machine learning model suitable for the prediction task is selected, such as random forest, neural network, etc. Multiple groups of training samples are divided into training set and test set, and the ratio can be 7:3 or 8:2. Using the training set data, basic information data and historical fund payment data are input, and the goal is to predict the fund loss rate. The prediction performance is optimized by adjusting the model parameters and hyperparameters, and a fund loss prediction model is trained.

[0045] Collect the basic information data of the target customers and the latest fund transfer data. The target date is set as the date after the preset number of days after the transfer funds arrive. The target date can be adjusted at any time based on expert experience to more accurately predict the fund loss rate on a certain day. Input the target customer data into the trained fund loss prediction model and output the predicted fund loss rate.

[0046] Step S203: when the predicted capital loss rate is greater than or equal to the capital loss rate threshold, the target customer is determined as a capital loss customer.

[0047] In step S203, by analyzing historical data, the average churn rate of customers who have lost funds in the past, as well as customer behavior characteristics and fund flow patterns under different churn rates are understood. The fund churn rate threshold is determined based on the business objectives of the financial institution, such as maintaining the stability of the customer base and increasing customer fund retention.

[0048] Compare the predicted fund churn rate of the target customer predicted by the model with the set fund churn rate threshold. If the predicted fund churn rate is greater than or equal to the threshold, the customer is identified as a fund churn customer, and the business department needs to adopt corresponding retention strategies.

[0049] For example, the capital loss rate R can be calculated by the following formula:

[0050]

[0051] Among them, Out t It represents the outflow amount after the funds are received on behalf of the account t days later (i.e. the target date), and account represents the amount of funds to be paid. According to the experience of business experts, there is a polarization phenomenon in the loss of funds on behalf of the account, and the loss rate of funds on behalf of the account is 50% as the threshold of the loss rate of funds.

[0052] Step S204, determining a target fund retention strategy based on the basic information data and the fund payment data, and executing the target fund retention strategy for the fund loss customer.

[0053] In step S204, the basic information data and the fund transfer data are deeply analyzed to identify the specific characteristics of the fund loss customers. For example, information such as age, occupation, fund transfer amount, fund loss frequency and amount, etc., will help formulate more targeted retention strategies. Based on the identified customer characteristics, similar customer groups of the target customers are determined. The target fund retention strategy is determined from the fund retention strategies of similar customer groups, and the target fund retention strategy is output for the fund loss customers with a higher probability of loss.

[0054] The method for processing fund loss of a client provided in an embodiment of the present application is by obtaining basic information data of a target client and fund transfer data within a preset period, wherein the basic information data includes at least one of the following: identity information data, fund distribution data and user behavior data, and the fund transfer data includes at least one of the following: average transfer amount, maximum transfer amount and average loss amount; inputting the basic information data and the fund transfer data into a fund loss prediction model to obtain a predicted fund loss rate of the target client on a target date, wherein the fund loss prediction model is obtained by training multiple groups of training samples, each group of training samples includes basic information data, historical fund transfer data and fund loss rate, and the target date is a date after a preset number of days from when the target client obtains the funds for transfer; When the predicted capital churn rate is greater than or equal to the capital churn rate threshold, the target customer is determined as a capital churn customer; the target capital retention strategy is determined based on the basic information data and the fund payment data, and the target capital retention strategy is implemented for the capital churn customer. The customer's capital churn rate on the target date is predicted through the customer's basic information data and the fund payment data, and then the customer is evaluated whether he is a capital churn customer based on the capital churn rate, and then the target capital retention strategy is implemented for the capital churn customer, thereby achieving the purpose of improving the accuracy of predicting capital churn for fund payment customers, thereby achieving the technical effect of improving the success rate of recovering capital churn for fund payment customers, and then solving the technical problem that the prediction of capital churn for fund payment customers is not accurate enough, resulting in a high capital churn for fund payment customers.

[0055] In order to screen out capital loss customers who need to implement capital retention strategies, it is necessary to train a capital loss prediction model. Optionally, in the method for processing capital loss of agency customers provided in the embodiment of the present application, the capital loss prediction model is trained in the following way: determine the agency customer group, obtain basic information data and historical agency data of the agency customer group; determine the capital loss rate of each agency customer on the target date from the historical agency data; determine the basic information data, historical agency data and capital loss rate of each agency customer as a group of training samples, and obtain multiple groups of training samples for the agency customer group; train a machine learning model based on the multiple groups of training samples to obtain a capital loss prediction model.

[0056] In some examples, first, it is necessary to determine the target customer group based on the agency payment business data of the financial institution, such as customers with a stable agency payment record in the past n months. Collect the basic information data and historical fund agency payment data of each customer in the agency payment customer group. The target date is set to the preset number of days after the agency payment funds are obtained, and the fund loss rate of each customer at that time point is calculated based on historical data. The fund loss rate can be evaluated by calculating the ratio of funds retained in the account after the agency payment funds are received, that is, (agency payment funds-loss funds) / agency payment funds.

[0057] Combine the basic information data of each agency customer, historical fund payment data, and fund loss rate on the target date into a set of training samples. Clean the collected data to remove outliers and missing data. Perform feature selection to eliminate redundant or irrelevant features. Divide all training samples into training sets and validation sets, such as 70% of the training samples are used for training, and 30% of the training samples are used to verify and test model performance. Use the training set data to train the machine learning model to predict the fund loss rate. During the model training process, you can adjust the model parameters through cross-validation, or adjust the model parameters through grid search to find the optimal configuration. Use the validation set to evaluate the model's key performance indicators such as prediction accuracy and recall rate. Based on the evaluation results, adjust the model parameters or improve feature engineering to optimize the model performance. Use the trained fund loss prediction model to predict the fund loss rate of new customers or agency customers at future time points. According to actual business feedback, continue to collect data, optimize the model, and form a closed-loop iteration mechanism.

[0058] By training a fund loss prediction model, we can timely predict the fund loss of the agency payment customers, mine each agency payment data, and predict whether the corresponding customer will be lost within a preset number of days. This helps financial institutions to timely identify high-risk customers and adopt corresponding retention strategies, thereby effectively reducing fund loss, improving customer satisfaction and the financial stability of the bank.

[0059] After determining the customers who have lost funds, similar customer groups are determined through basic information data and fund payment data, and then a target fund retention strategy is determined based on the fund retention strategies of the similar customer groups. Optionally, in the method for handling fund loss of fund payment customers provided in the embodiment of the present application, determining the target fund retention strategy based on basic information data and fund payment data includes: determining similar customer groups of target customers based on basic information data and fund payment data; obtaining multiple fund retention strategies executed by similar customer groups within a target period, and determining the retention success rate of each fund retention strategy to obtain a set of retention success rates; and determining the fund retention strategy corresponding to the maximum retention success rate in a set of retention success rates as the target fund retention strategy.

[0060] In some examples, first, using basic information data and fund distribution data, similarity calculation methods such as the Pearson correlation coefficient are used to identify customers who are similar to the target customers in terms of behavior habits, fund management habits, product preferences, etc. These customers constitute a similar customer group. The fund retention strategies implemented by similar customer groups during the target period (such as the previous fund churn rate prediction period) are collected, including but not limited to product recommendations, equity rewards, customer service upgrades, etc. At the same time, the execution results of each strategy, that is, the retention success rate, are recorded.

[0061] The retention success rate can be calculated by the ratio of the number of customers recovered / the total number of customers marked as losing funds, where the number of customers recovered refers to the number of customers whose fund loss rate is significantly reduced or whose fund retention is increased after implementing a specific retention strategy; the total number of customers marked as losing funds refers to the total number of customers who are predicted to lose funds within the target period. Based on the collected fund retention strategies and their corresponding retention success rates, select the strategy with the highest retention success rate as the target fund retention strategy. Apply the selected retention strategy to the target customers.

[0062] Among them, the fund retention strategy can include 1) offline direct contact, public-private linkage fund retention strategy: Since the agency business takes into account both corporate and individual customers, starting from public enterprises, the individual customer churn in the predicted results will be classified into public units. Business personnel can carry out offline workplace product recommendations for agencies with high churn rates or institutional agency customers, key industry agency customers and other agencies whose agency customers have stable jobs and considerable incomes, and communicate with the agency units and their employees face to face. Carry out targeted activities such as "lucky draw gifts", "card binding gifts", and "card application gifts", and guide branches to strengthen the penetration of basic products and agency-exclusive products through financial product manuals. At the same time, establish a customer feedback mechanism to collect customer suggestions and opinions in a timely manner, make improvements, and continuously optimize service quality and improve customer satisfaction.

[0063] 2) Improve product penetration and enhance customer stickiness: For frequent transfers from accounts with the same name, or large amounts of funds transferred across banks, it can be inferred that customers are likely to transfer their wages immediately after receiving them. These customers may have large funds but insufficient customer loyalty. For this type of customer group, improving customer product penetration is an important means to reduce capital outflows. It is necessary to use products and customer rights as touchpoints, continuously maintain the work through corporate micro-blogs, outbound calls, etc., and recommend products such as third-party card binding, credit cards, and loans to customers.

[0064] 3) Linking mobile banking to recommend financial products: For customers with high agency payout and low retention, such as customers with agency payout amount of more than 10,000 but monthly average assets of less than 10,000, recommending financial products to them is a major way to reduce capital outflow. According to the experience of business experts, in view of the characteristics of the overall low transaction amount and higher transaction frequency of the young group, asset management products with lower transaction starting point and stronger liquidity can be matched, such as open-end funds, short-term profit-holding financial products, etc. For middle-aged and elderly customers who are slightly older, the allocation strategy in terms of liquidity can be reduced, and the focus can be on the stability of the customer group's long-term holding assets, such as multi-year term insurance, regular open-end financial products, etc.

[0065] This embodiment can not only predict customers who are likely to lose funds by determining the target fund retention strategy, but also design a personalized and efficient retention strategy for each target customer based on the historical data of similar customer groups and the effectiveness of the strategy, thereby effectively reducing the fund loss rate and improving customer retention and satisfaction.

[0066] Similar customer groups can be determined by the similarity between the basic information data and the fund payment data of each customer. Optionally, in the method for handling the loss of funds of the fund payment customer provided in the embodiment of the present application, determining the similar customer groups of the target customer based on the basic information data and the fund payment data includes: obtaining the basic information data and the fund payment data of each fund payment customer, converting the basic information data and the fund payment data of each fund payment customer into a first feature vector to obtain a set of first feature vectors; converting the basic information data and the fund payment data of the target customer into a second feature vector; calculating the similarity between the second feature vector and each first feature vector; determining the fund payment customer corresponding to the first feature vector whose similarity is greater than or equal to the similarity threshold as a similar customer, and defining the set of all identified similar customers as a similar customer group.

[0067] In some examples, first, basic information data and fund transfer data of each agency customer are collected, and the basic information data and fund transfer data of each agency customer are converted into a first feature vector represented by a numerical value. For example, age may be converted into a numerical value, occupation may be encoded as a digital identifier, and the agency amount directly uses a specific numerical value. Similarly, the basic information data and fund transfer data of the target customer are converted into a second feature vector.

[0068] The similarity between the second eigenvector and each first eigenvector is calculated using methods such as Pearson correlation coefficient and cosine similarity. The higher the similarity, the greater the similarity between the two customers in terms of behavior patterns and fund management habits. A similarity threshold is set to distinguish which customers can be considered as "similar customers" of the target customers.

[0069] If the similarity between the second feature vector and a first feature vector is greater than or equal to the similarity threshold, the agency customer represented by the corresponding first feature vector is regarded as a similar customer of the target customer. The set of all agency customers whose similarity with the target customer is greater than or equal to the threshold constitutes a similar customer group of the target customer.

[0070] This embodiment identifies similar customer groups with similar characteristics and behavioral habits as the target customers, providing a data basis for subsequent fund retention strategy recommendations based on similar customer groups, helping to improve the effectiveness and pertinence of the strategy, thereby reducing capital loss and increasing customer loyalty.

[0071] For customers who have suffered capital loss after the target capital retention strategy has been implemented, the retention success rate of the target capital retention strategy is updated according to the retention results. Optionally, in the method for handling capital loss of payee customers provided in the embodiment of the present application, after implementing the target capital retention strategy on the customers who have suffered capital loss, the method further includes: determining the actual capital loss rate of the target customers on the target date, and judging whether the actual capital loss rate is greater than or equal to a capital loss rate threshold; determining that the target capital retention strategy has failed to retain the customers if the actual capital loss rate is greater than or equal to the capital loss rate threshold; determining that the target capital retention strategy has succeeded to retain the customers if the actual capital loss rate is less than the capital loss rate threshold; and updating the retention success rate of the target capital retention strategy based on the retention results of the target capital retention strategy.

[0072] In some examples, after the target date, the target customer's fund flow information in the bank account is collected, including but not limited to transaction records such as transfers, cash withdrawals, and consumption. By analyzing the target customer's fund flow over a period of time after the target date, the actual amount of fund loss is determined, and then the actual fund loss rate is calculated. The calculated actual fund loss rate is compared with the set fund loss rate threshold. If the actual fund loss rate is greater than or equal to the fund loss rate threshold, it means that the customer has significant fund loss behavior; conversely, if the actual fund loss rate is less than the threshold, it indicates that the fund loss is effectively controlled.

[0073] If the actual fund loss rate is greater than or equal to the fund loss rate threshold, it means that although the target fund retention strategy has been implemented, it has not been able to effectively prevent fund loss, and the retention strategy is considered a failure. On the contrary, if the actual fund loss rate is lower than the threshold, it means that the retention strategy has successfully prevented or reduced fund loss, and the strategy is considered a success. Record the execution results of the retention strategy for each target customer, that is, mark it as success or failure. Based on the recorded retention strategy execution results, recalculate the retention success rate of each retention strategy. Adjust the priority and execution details of the fund retention strategy based on the updated retention success rate.

[0074] This embodiment provides customers with more accurate fund retention strategies by continuously optimizing the retention success rate of the fund retention strategies, effectively reducing fund loss and enhancing customer loyalty.

[0075] In order to improve the accuracy and real-time performance of the capital loss prediction model, the capital loss prediction model is updated regularly. Optionally, in the method for processing capital loss of a client provided in an embodiment of the present application, after obtaining the predicted capital loss rate of the target client on the target date, the method further includes: when the predicted capital loss rate is less than a capital loss rate threshold, determining the actual capital loss rate of the target client on the target date, and judging whether the actual capital loss rate is greater than or equal to the capital loss rate threshold; when the actual capital loss rate is less than the capital loss rate threshold, determining that the capital loss rate prediction result of the target client is accurate; when the actual capital loss rate is greater than or equal to the capital loss rate threshold, determining that the capital loss rate prediction result of the target client is erroneous; and calculating the accuracy of the capital loss rate prediction result of the capital loss prediction model within the statistical model update cycle, and updating the capital loss prediction model when the accuracy is less than the accuracy threshold.

[0076] In some examples, if the predicted fund churn rate is less than the set fund churn rate threshold, it means that the model predicts that the customer will not have significant fund churn on the target date. After the target date, the actual fund flow information of the target customer after that date is collected to calculate the actual fund churn rate. If the actual fund churn rate is less than the fund churn rate threshold, this is consistent with the model's prediction result, indicating that the model successfully predicted that the customer will not have significant fund churn. Therefore, the prediction result is determined to be accurate.

[0077] If the actual fund loss rate is greater than or equal to the fund loss rate threshold, this is inconsistent with the model prediction result, indicating that the model failed to accurately predict that the customer will lose funds. Therefore, the prediction result is determined to be wrong. During the model update cycle, all prediction results are counted and the model prediction accuracy is calculated. The accuracy rate is the ratio of customers who actually did not experience significant fund loss among all customers who were predicted by the model to not experience significant fund loss during the prediction cycle. Set an accuracy threshold. If the fund loss prediction accuracy obtained during the model update cycle is less than the set accuracy threshold, it means that the model performs poorly and needs to be updated to improve the prediction ability.

[0078] The update of the fund churn prediction model can add the latest customer information and fund flow data to the training data set to reflect recent behavioral trends. Retrain the model with the updated training data set and adjust the model parameters to improve the prediction accuracy. Re-evaluate the model performance on the validation set to ensure that the new model performs better than the previous model in predicting fund churn rate.

[0079] This embodiment ensures that the capital loss prediction model can adapt to the ever-changing market environment and customer behavior by continuously monitoring the prediction accuracy of the model and making timely adjustments, thereby more effectively preventing capital loss, maintaining customer relationships, and improving the overall business performance and risk management capabilities of financial institutions.

[0080] Even if the actual capital loss rate is less than the capital loss rate threshold, if the difference between the predicted capital loss rate and the actual capital loss rate is too large, the capital loss rate prediction result of the target customer is still determined to be wrong. Optionally, in the method for processing capital loss of a payee provided in an embodiment of the present application, before determining that the capital loss rate prediction result of the target customer is accurate, the method further includes: calculating the difference between the predicted capital loss rate and the actual capital loss rate, and determining whether the difference is greater than or equal to a preset numerical threshold; if the difference is greater than or equal to the preset numerical threshold, determining that the capital loss rate prediction result of the target customer is wrong; if the difference is less than the preset numerical threshold, determining that the capital loss rate prediction result of the target customer is accurate.

[0081] In some examples, the predicted churn rate is obtained from the model output, and the actual churn rate is calculated from the customer account data after the target date. The difference between the predicted churn rate and the actual churn rate is calculated and compared with a preset numerical threshold. If the difference is greater than or equal to the preset numerical threshold, it is determined that the prediction result of the target customer's churn rate is wrong, indicating that the deviation between the model-predicted churn rate and the actual churn rate is too large, and the expected prediction accuracy is not achieved.

[0082] If the difference is less than the preset numerical threshold, it is determined that the prediction result of the target customer's capital churn rate is accurate, indicating that the prediction of the model is consistent with the actual situation and the prediction accuracy is within an acceptable range.

[0083] This embodiment continuously monitors and evaluates the performance of the fund loss prediction model to ensure that the model can accurately predict the loss of customer funds, so that effective measures can be taken in a timely manner to reduce fund loss and maintain customer relationships.

[0084] According to another embodiment of the present application, an optional method for handling the loss of funds of a client who is responsible for sending funds is also provided. Figure 3 is a flowchart of an optional method for processing the loss of funds of a client provided by an embodiment of the present application, such as Figure 3As shown, the method includes: first, obtaining the original customer data of the agency customer within the set time window, including basic information, asset distribution, behavioral habits, product stickiness and other features. Performing data cleaning, missing value processing and feature vectorization on the original customer. After dividing the agency customer feature data set after feature screening, constructing an agency customer fund loss prediction model and outputting the prediction results. After the prediction is completed, the agency customer fund retention strategy recommendation algorithm is applied to output the recovery strategy for the agency customers with a higher probability of loss, and the data with better strategy execution effect is further fed back to the algorithm data set for data update.

[0085] Among them, the process of the recommended algorithm for the client funds retention strategy is as follows: Figure 4 As shown, Figure 4 is a schematic diagram of the execution process of the client funds retention strategy recommendation algorithm provided in the embodiment of the present application, such as Figure 4 As shown in the figure, based on the feature data set of the agency payment customers and the prediction results of the agency payment customers' fund loss, the feature data set for the recommendation of the agency payment customer fund retention strategy is constructed, the user-based collaborative filtering algorithm is used to calculate the optimal agency payment customer fund retention strategy, and the Peason (Pearson similarity algorithm) coefficient is used to determine the similar customers of the agency payment customers. Provide accurate customer retention strategy recommendations for agency payment customers, and the business department executes the strategy to retain funds for the agency payment customers.

[0086] This embodiment can timely predict the fund loss of the agency customer through the optional method of processing the fund loss of the agency customer, mine each agency data, and predict whether the corresponding customer will be lost within a preset number of days. The model supports aggregation to the agency enterprise dimension, reminding the business department to adopt product recommendation methods and retention strategies before the agency funds are lost, and accurately recommends corresponding retention strategies based on historical data, effectively improving the bank's agency business maintenance efficiency. It has high accuracy in predicting customer fund loss rate, providing business departments with agency fund loss reminders accurate to a certain day, and the output of the strategy recommendation model directly assists the salesperson to take corresponding measures in a timely manner. It has good scalability, can adapt to different agency business scenarios, and improve the application effect of the model.

[0087] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0088] Example 2

[0089] The embodiment of the present application also provides a processing device for the loss of funds of an agent-sending client. It should be noted that the processing device for the loss of funds of an agent-sending client in the embodiment of the present application can be used to execute the processing method for the loss of funds of an agent-sending client provided in the embodiment of the present application. The following introduces the processing device for the loss of funds of an agent-sending client provided in the embodiment of the present application.

[0090] According to an embodiment of the present application, a device for implementing the above-mentioned method for processing the loss of funds of a client who is to be paid on behalf of another is also provided. Figure 5 is a schematic diagram of a device for processing fund loss of a client provided by an embodiment of the present application, such as Figure 5 As shown, the device comprises:

[0091] The acquisition unit 501 is used to acquire basic information data of the target customer and fund transfer data within a preset period, wherein the basic information data includes at least one of the following: identity information data, fund distribution data and user behavior data, and the fund transfer data includes at least one of the following: average transfer amount, maximum transfer amount and average loss amount;

[0092] An input unit 502 is used to input the basic information data and the fund payment data into a fund loss prediction model to obtain a predicted fund loss rate of a target customer on a target date, wherein the fund loss prediction model is trained by a plurality of training samples, each training sample comprising basic information data, historical fund payment data and fund loss rate, and the target date is a date after a preset number of days from when the target customer obtains the payment funds;

[0093] A first determining unit 503 is used to determine the target customer as a capital loss customer when the predicted capital loss rate is greater than or equal to a capital loss rate threshold;

[0094] The execution unit 504 is used to determine a target fund retention strategy based on the basic information data and the fund payment data, and execute the target fund retention strategy for the fund loss customer.

[0095] The processing device for fund loss of a client provided in the embodiment of the present application is used to obtain basic information data of a target client and fund transfer data within a preset period through an acquisition unit 501, wherein the basic information data includes at least one of the following: identity information data, fund distribution data and user behavior data, and the fund transfer data includes at least one of the following: average transfer amount, maximum transfer amount and average loss amount; an input unit 502 is used to input the basic information data and fund transfer data into a fund loss prediction model to obtain a predicted fund loss rate of the target client on a target date, wherein the fund loss prediction model is obtained by training multiple groups of training samples, each group of training samples includes basic information data, historical fund transfer data and fund loss rate, and the target date is a date after a preset number of days from when the target client obtains the funds for transfer; A determination unit 503 is used to determine the target customer as a capital loss customer when the predicted capital loss rate is greater than or equal to a capital loss rate threshold; an execution unit 504 is used to determine a target capital retention strategy based on basic information data and fund payment data, and execute the target capital retention strategy on the capital loss customer. The capital loss rate of the customer on the target date is predicted through the customer's basic information data and fund payment data, and then the customer is evaluated based on the capital loss rate to see whether the customer is a capital loss customer, and then the target capital retention strategy is executed on the capital loss customer, thereby achieving the purpose of improving the accuracy of predicting capital loss of payment customers, thereby achieving the technical effect of improving the success rate of recovering capital loss of payment customers, and then solving the technical problem that the prediction of capital loss of payment customers is not accurate enough, resulting in a high capital loss of payment customers.

[0096] Optionally, in the device for processing fund loss of agency customers provided in the embodiment of the present application, the device also includes: a second determination unit, used to determine the agency customer group, and obtain basic information data and historical fund payment data of the agency customer group; a third determination unit, used to determine the fund loss rate of each agency customer on the target date from the historical fund payment data; a fourth determination unit, used to determine the basic information data, historical fund payment data and fund loss rate of each agency customer as a group of training samples, and obtain multiple groups of training samples for the agency customer group; a training unit, used to train a machine learning model based on the multiple groups of training samples to obtain a fund loss prediction model.

[0097] Optionally, in the processing device for fund loss of agency customers provided in the embodiment of the present application, the execution unit 504 includes: a first determination module, used to determine similar customer groups of target customers based on basic information data and fund payment data; an acquisition module, used to obtain multiple fund retention strategies executed by similar customer groups within a target period, and determine the retention success rate of each fund retention strategy to obtain a group of retention success rates; a second determination module, used to determine the fund retention strategy corresponding to the maximum retention success rate in a group of retention success rates as the target fund retention strategy.

[0098] Optionally, in the processing device for fund loss of agency customers provided in the embodiment of the present application, the first determination module includes: an acquisition submodule, used to acquire basic information data and fund payment data of each agency customer, and convert the basic information data and fund payment data of each agency customer into a first feature vector to obtain a group of first feature vectors; a conversion submodule, used to convert the basic information data and fund payment data of the target customer into a second feature vector; a calculation submodule, used to calculate the similarity between the second feature vector and each first feature vector; and a determination submodule, used to determine the agency customer corresponding to the first feature vector whose similarity is greater than or equal to the similarity threshold as a similar customer, and define the set of all identified similar customers as a similar customer group.

[0099] Optionally, in the device for processing fund loss of an agent customer provided in an embodiment of the present application, the device also includes: a fifth determination unit, used to determine the actual fund loss rate of the target customer on the target date, and to determine whether the actual fund loss rate is greater than or equal to a fund loss rate threshold; a sixth determination unit, used to determine that the target fund retention strategy has failed to retain the target fund when the actual fund loss rate is greater than or equal to the fund loss rate threshold; a seventh determination unit, used to determine that the target fund retention strategy has succeeded to retain the target fund when the actual fund loss rate is less than the fund loss rate threshold; and a first update unit, used to update the retention success rate of the target fund retention strategy based on the retention result of the target fund retention strategy.

[0100] Optionally, in the processing device for client funds loss provided in the embodiment of the present application, the device also includes: an eighth determination unit, for determining the actual funds loss rate of the target customer on the target date when the predicted funds loss rate is less than the funds loss rate threshold, and judging whether the actual funds loss rate is greater than or equal to the funds loss rate threshold; a ninth determination unit, for determining that the funds loss rate prediction result of the target customer is accurate when the actual funds loss rate is less than the funds loss rate threshold; a tenth determination unit, for determining that the funds loss rate prediction result of the target customer is wrong when the actual funds loss rate is greater than or equal to the funds loss rate threshold; a second updating unit, for statistically calculating the accuracy of the funds loss rate prediction result of the funds loss prediction model within the model update period, and updating the funds loss prediction model when the accuracy is less than the accuracy threshold.

[0101] Optionally, in the processing device for processing fund loss of an agent customer provided in an embodiment of the present application, the device also includes: a calculation unit, used to calculate the difference between the predicted fund loss rate and the actual fund loss rate, and determine whether the difference is greater than or equal to a preset numerical threshold; an eleventh determination unit, used to determine that the predicted result of the fund loss rate of the target customer is wrong when the difference is greater than or equal to the preset numerical threshold; a twelfth determination unit, used to determine that the predicted result of the fund loss rate of the target customer is accurate when the difference is less than the preset numerical threshold.

[0102] It should be noted that the acquisition unit 501, input unit 502, first determination unit 503 and execution unit 504 correspond to steps S201 to S204 in Example 1, and the four units and corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above-mentioned Example 1. It should be noted that the above-mentioned modules or units may be hardware components or software components stored in a memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n), and the above-mentioned modules or units may also be part of a device and may be run in the computer terminal 10 provided in Example 1.

[0103] Example 3

[0104] An embodiment of the present application may provide an electronic device, Figure 6 is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 6 As shown, the electronic device may include: one or more ( Figure 6 (only one is shown) processor 602, memory 604, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0105] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0106] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: obtain the basic information data of the target customer and the fund payment data within a preset period, wherein the basic information data includes at least one of the following: identity information data, fund distribution data and user behavior data, and the fund payment data includes at least one of the following: average payment amount, maximum payment amount and average loss amount; input the basic information data and the fund payment data into the fund loss prediction model to obtain the predicted fund loss rate of the target customer on the target date, wherein the fund loss prediction model is trained by multiple groups of training samples, each group of training samples includes basic information data, historical fund payment data and fund loss rate, and the target date is the date after the preset number of days when the target customer obtains the payment funds; when the predicted fund loss rate is greater than or equal to the fund loss rate threshold, the target customer is determined as a fund loss customer; based on the basic information data and the fund payment data, a target fund retention strategy is determined, and the target fund retention strategy is executed on the fund loss customer.

[0107] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: determine the agency customer group, obtain the basic information data and historical fund payment data of the agency customer group; determine the fund loss rate of each agency customer on the target date from the historical fund payment data; determine the basic information data, historical fund payment data and fund loss rate of each agency customer as a group of training samples to obtain multiple groups of training samples for the agency customer group; train the machine learning model based on the multiple groups of training samples to obtain a fund loss prediction model.

[0108] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: determine similar customer groups of target customers based on basic information data and fund payment data; obtain multiple fund retention strategies executed by similar customer groups within the target period, and determine the retention success rate of each fund retention strategy to obtain a set of retention success rates; determine the fund retention strategy corresponding to the maximum retention success rate in a set of retention success rates as the target fund retention strategy.

[0109] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: obtain the basic information data and fund payment data of each agency customer, convert the basic information data and fund payment data of each agency customer into a first feature vector, and obtain a set of first feature vectors; convert the basic information data and fund payment data of the target customer into a second feature vector; calculate the similarity between the second feature vector and each first feature vector; determine the agency customer corresponding to the first feature vector whose similarity is greater than or equal to the similarity threshold as a similar customer, and define the set of all identified similar customers as a similar customer group.

[0110] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: determine the actual funds loss rate of the target customer on the target date, and determine whether the actual funds loss rate is greater than or equal to the funds loss rate threshold; when the actual funds loss rate is greater than or equal to the funds loss rate threshold, determine that the target funds retention strategy has failed to retain; when the actual funds loss rate is less than the funds loss rate threshold, determine that the target funds retention strategy has succeeded in retaining; and update the retention success rate of the target funds retention strategy based on the retention result of the target funds retention strategy.

[0111] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: when the predicted capital loss rate is less than the capital loss rate threshold, determine the actual capital loss rate of the target customer on the target date, and judge whether the actual capital loss rate is greater than or equal to the capital loss rate threshold; when the actual capital loss rate is less than the capital loss rate threshold, determine that the capital loss rate prediction result of the target customer is accurate; when the actual capital loss rate is greater than or equal to the capital loss rate threshold, determine that the capital loss rate prediction result of the target customer is wrong; and calculate the accuracy of the capital loss rate prediction result of the capital loss prediction model within the statistical model update period, and when the accuracy is less than the accuracy threshold, update the capital loss prediction model.

[0112] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: calculate the difference between the predicted capital loss rate and the actual capital loss rate, and determine whether the difference is greater than or equal to a preset numerical threshold; when the difference is greater than or equal to the preset numerical threshold, determine that the predicted result of the capital loss rate of the target customer is wrong; when the difference is less than the preset numerical threshold, determine that the predicted result of the capital loss rate of the target customer is accurate.

[0113] By adopting the embodiment of the present application, a method is provided for obtaining basic information data of a target customer and fund payment data within a preset period, wherein the basic information data includes at least one of the following: identity information data, fund distribution data and user behavior data, and the fund payment data includes at least one of the following: average payment amount, maximum payment amount and average loss amount; the basic information data and the fund payment data are input into a fund loss prediction model to obtain a predicted fund loss rate of the target customer on a target date, wherein the fund loss prediction model is trained by multiple groups of training samples, each group of training samples includes basic information data, historical fund payment data and fund loss rate, and the target date is a date after a preset number of days from when the target customer obtains the payment funds; when the predicted fund loss rate is greater than or equal to a fund loss rate threshold, the target customer is determined as a fund loss customer; a target fund retention strategy is determined based on the basic information data and the fund payment data, and a plan for executing the target fund retention strategy on the fund loss customer is implemented. The customer's fund loss rate on the target date is predicted by using the customer's basic information data and fund payment data, and then the customer is evaluated whether he or she is a fund loss customer based on the fund loss rate, and then a target fund retention strategy is implemented for the fund loss customer, thereby achieving the purpose of improving the prediction accuracy of fund loss for fund payment customers, thereby achieving the technical effect of improving the success rate of recovering fund loss for fund payment customers, and thus solving the technical problem that the prediction of fund loss for fund payment customers is not accurate enough, resulting in a high fund loss for fund payment customers.

[0114] It can be understood by those skilled in the art that Figure 6 The structure shown is for illustration only, and the electronic device may also be a smart phone, a tablet computer, a PDA, a mobile Internet device (MID), a PAD or other terminal device. Figure 6 The structure of the electronic device is not limited. Figure 6 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 6 Different configurations are shown.

[0115] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0116] Example 4

[0117] The embodiment of the present application further provides a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the method for processing the loss of funds of the agency-sending client provided in the first embodiment.

[0118] Optionally, in this embodiment, the above storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0119] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing the program steps of the method for processing the loss of funds of a proxy customer.

[0120] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0121] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0122] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0123] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0124] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0125] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0126] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for handling the loss of funds of a client who is responsible for sending money, characterized in that: include: Obtaining basic information data of target customers and fund transfer data within a preset period, wherein the basic information data includes at least one of the following: identity information data, fund distribution data, and user behavior data, and the fund transfer data includes at least one of the following: average transfer amount, maximum transfer amount, and average loss amount; Inputting the basic information data and the fund payment data into a fund loss prediction model to obtain a predicted fund loss rate of the target customer on a target date, wherein the fund loss prediction model is trained by multiple sets of training samples, each set of training samples includes basic information data, historical fund payment data and fund loss rate, and the target date is a date after a preset number of days from when the target customer obtains the payment funds; When the predicted capital loss rate is greater than or equal to a capital loss rate threshold, determining the target customer as a capital loss customer; A target fund retention strategy is determined based on the basic information data and the fund payment data, and the target fund retention strategy is executed on the fund-losing customer.

2. The method according to claim 1, characterized in that The fund loss prediction model is trained in the following way: Determine the client group for fund transfer, and obtain basic information data and historical fund transfer data of the client group for fund transfer; Determine the fund loss rate of each fund transfer client on the target date from the historical fund transfer data; Determine the basic information data, historical fund payment data and fund loss rate of each agency payment customer as a group of training samples, and obtain multiple groups of training samples of the agency payment customer group; The machine learning model is trained based on the multiple groups of training samples to obtain the capital loss prediction model.

3. The method according to claim 1, characterized in that Determining the target fund retention strategy based on the basic information data and the fund payment data includes: Determine a similar customer group to the target customer based on the basic information data and the fund payment data; Acquire multiple fund retention strategies executed by the similar customer group within the target period, and determine the retention success rate of each fund retention strategy to obtain a set of retention success rates; The fund retention strategy corresponding to the maximum retention success rate among the group of retention success rates is determined as the target fund retention strategy.

4. The method according to claim 3, characterized in that Determining similar customer groups of the target customer based on the basic information data and the fund payment data includes: Obtaining basic information data and fund payment data of each agency payment customer, converting the basic information data and fund payment data of each agency payment customer into a first feature vector, and obtaining a set of first feature vectors; Converting the target customer's basic information data and fund transfer data into a second feature vector; Calculating the similarity between the second feature vector and each first feature vector; The sending customers corresponding to the first feature vectors whose similarity is greater than or equal to the similarity threshold are determined as similar customers, and the set of all identified similar customers is defined as the similar customer group.

5. The method according to claim 1, characterized in that After executing the target funds retention strategy for the funds-losing customer, the method further includes: Determine the actual capital loss rate of the target customer on the target date, and determine whether the actual capital loss rate is greater than or equal to the capital loss rate threshold; When the actual fund loss rate is greater than or equal to the fund loss rate threshold, determining that the target fund retention strategy has failed to retain the target fund; When the actual capital loss rate is less than the capital loss rate threshold, determining that the target capital retention strategy is successful in retaining capital; The retention success rate of the target funds retention strategy is updated based on the retention result of the target funds retention strategy.

6. The method according to claim 1, characterized in that After obtaining the predicted capital churn rate of the target customer on the target date, the method further includes: In the case where the predicted capital loss rate is less than the capital loss rate threshold, determining the actual capital loss rate of the target customer on the target date, and judging whether the actual capital loss rate is greater than or equal to the capital loss rate threshold; When the actual capital loss rate is less than the capital loss rate threshold, determining that the capital loss rate prediction result of the target customer is accurate; When the actual capital loss rate is greater than or equal to the capital loss rate threshold, determining that the capital loss rate prediction result of the target customer is wrong; The accuracy of the fund loss rate prediction result of the fund loss prediction model within the statistical model update period is calculated, and when the accuracy is less than the accuracy threshold, the fund loss prediction model is updated.

7. The method according to claim 6, characterized in that Before determining that the target customer's fund churn rate prediction result is accurate, the method further includes: Calculating the difference between the predicted capital loss rate and the actual capital loss rate, and determining whether the difference is greater than or equal to a preset numerical threshold; When the difference is greater than or equal to the preset numerical threshold, determining that the prediction result of the target customer's capital loss rate is wrong; When the difference is less than the preset numerical threshold, it is determined that the prediction result of the target customer's capital loss rate is accurate.

8. A device for processing the loss of funds of a client who is sending money on behalf of another, characterized in that: include: An acquisition unit, used to acquire basic information data of a target customer and fund transfer data within a preset period, wherein the basic information data includes at least one of the following: identity information data, fund distribution data, and user behavior data, and the fund transfer data includes at least one of the following: average transfer amount, maximum transfer amount, and average loss amount; An input unit, used for inputting the basic information data and the fund payment data into a fund loss prediction model to obtain a predicted fund loss rate of the target customer on a target date, wherein the fund loss prediction model is obtained by training multiple groups of training samples, each group of training samples includes basic information data, historical fund payment data and fund loss rate, and the target date is a date after a preset number of days for the target customer to obtain the payment funds; A first determining unit, configured to determine the target customer as a capital loss customer when the predicted capital loss rate is greater than or equal to a capital loss rate threshold; An execution unit is used to determine a target fund retention strategy based on the basic information data and the fund payment data, and execute the target fund retention strategy for the fund loss customer.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the method for processing the loss of funds of the agency customer as described in any one of claims 1 to 7.

10. An electronic device, characterized in that: include: A memory storing an executable program; A processor is used to run the program, wherein when the program is running, the method for processing the loss of funds of the agency-paying customer described in any one of claims 1 to 7 is executed.

11. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by the processor, the steps of the method for processing the loss of funds of the agency-paying customer described in any one of claims 1 to 7 are implemented.

Images