CN112541817A

CN112541817A - Marketing response processing method and system for potential customers of personal consumption loan

Info

Publication number: CN112541817A
Application number: CN202011526366.4A
Authority: CN
Inventors: 陈芷君; 张晴
Original assignee: CCB Finetech Co Ltd
Current assignee: China Construction Bank Corp
Priority date: 2020-12-22
Filing date: 2020-12-22
Publication date: 2021-03-23

Abstract

The invention discloses a marketing response processing method and a marketing response processing system for potential customers of personal consumption loan, wherein the method comprises the following steps: screening potential customers according to customer data, and calculating the pre-granted loan amount in batches; selecting client data recorded with personal consumption loan amount measuring and calculating behaviors in a historical time range as a model training sample, and performing machine learning training by taking whether a client obtaining the amount measuring and calculating subsequently applies for loan as a target variable or not, and establishing a client response probability prediction model; and according to the client data of the potential client and the pre-credit loan amount, predicting the probability of applying for loan after the client knows the measuring and calculating amount by using a client response probability prediction model, grading the potential client according to the probability, and adopting a corresponding marketing strategy according to the grading result. The invention can provide more easily understood and operated basis for the business development of marketing personnel, realize differentiated and accurate marketing and effectively improve the marketing efficiency and the conversion rate.

Description

Marketing response processing method and system for potential customers of personal consumption loan

Technical Field

The invention relates to the technical field of internet finance, in particular to a marketing response processing method and system for potential customers of personal consumption loan.

Background

In recent years, with the increase of national consumption level and the transformation of consumption concept, the scale of domestic consumption credit market is growing at a high speed. The construction and perfection of credit investigation system and the improvement of credit investigation coverage rate also promote the development of consumption credit in China. However, in the face of the diversified current situation that the economy is accelerated and slowed down, and the internet financial institutions of non-banking systems participate in market competition, the traditional banking industry needs to adjust the business development strategy to seize the market share of personal consumption loan and better conduct the business of personal consumption loan, and the new technology is utilized to fully exert the advantages of the business development strategy to realize the featured fine operation.

The stock customers are important resources of banks, the cardinality is huge, and stable product customer growth can be obtained if effective personal consumption loan product marketing can be carried out on the stock customers, and conversion from potential customers to product formal customers is realized. At present, digital marketing is proposed in the prior art, and a client image and a client tag are established from the aspects of client asset liability, risk, consumption behavior, social attributes and the like by constructing a client tag system so as to assist marketing personnel in developing business. Although the customer group of the customer can be divided to a certain extent, the demand direction of the customer is approximately provided, the analysis on the application demand and the intention of the customer about a specific loan product is lacked, the possible response degree of the customer to marketing is not predicted, the marketing effect depends on the understanding of a marketing person to a label, the intention of the customer cannot be effectively predicted and distinguished, so that the accurate differentiated marketing cannot be carried out, and the requirement of improving the marketing conversion rate and reducing the marketing cost cannot be met.

Therefore, a marketing response technical solution that can overcome the above problems and improve marketing conversion rate while reducing marketing cost is needed.

Disclosure of Invention

In order to solve the problems in the prior art, the invention provides a marketing response processing method and system for a potential customer of personal consumption loan, the method and system can establish a response model for the potential customer of personal consumption loan, the customer is graded according to the prediction score of the response model, the application probability levels of different grades are different, and then business personnel can implement differentiated marketing strategies for the customers with different application probability levels, so that the marketing conversion rate is improved, and more accurate and effective marketing is realized.

In a first aspect of an embodiment of the present invention, a marketing response processing method for a potential customer of a personal consumption loan is provided, the method includes:

screening potential customers according to customer data, and calculating corresponding pre-credit loan amounts in batches;

in the client data, selecting client data recorded with personal consumption loan amount measuring and calculating behaviors in a historical time range as a model training sample, performing machine learning training by taking whether a client obtaining the measured and calculated amount subsequently applies for loan as a target variable, and establishing a client response probability prediction model which is used for predicting the probability of applying for loan after the client knows the measured and calculated amount;

and predicting the probability of applying for loan after the client knows the measuring and calculating amount by using the client response probability prediction model according to the client data of the potential client and the pre-credit loan amount, grading the potential client according to the probability, and adopting a corresponding marketing strategy according to the grading result.

Further, the customer data includes at least: client contribution degree, proxy wage information, credit card information, inline loan information, personal consumption loan amount measurement and calculation information and application behavior information of the client.

Further, screening potential customers according to the customer data, and calculating corresponding pre-credit loan amounts in batches, wherein the steps comprise:

determining the range of the individual loan product consumption according to the service requirement, the application scene and the data availability;

screening a first customer group which accords with the conditions of a public admission rule in the customer data according to the range of personal consumption loan products;

according to the customer data of the first customer group, removing customers with inventory unrefined personal consumption loan accounts to obtain a second customer group;

and carrying out pre-granted credit limit measurement and calculation according to the client data of the second client group, and reserving the client with the measurement and calculation result larger than 0 as a potential client.

Further, the common admission rule conditions include at least: non-blacklist customers, non-grey list customers, anti-money laundering verification pass, no loan overdue exists, non-internal control list customers and are in line with age restrictions.

Further, the method for measuring and calculating the pre-granted credit limit according to the client data of the second client group and reserving the client with the measuring and calculating result larger than 0 as a potential client comprises the following steps:

and pre-granting credit line budget according to the client contribution degree, the proxy wage information and the house property information of the second client group.

Further, the method further comprises:

generating derived variables related to the business in batches according to the client data;

and according to the derived variables, primarily screening the derived variables through coverage rate inspection and population stability index calculation and information value and correlation inspection, and taking the screened derived variables as model training samples.

Further, generating business related derivative variables according to the client data batch comprises:

and aggregating and converting the original characteristics according to the client data to derive characteristic variables with service significance.

Further, aggregating the original features according to the customer data to derive feature variables with business significance, including:

and aggregating the statistics of various parameters of the original characteristics, and deriving characteristic variables with service significance after adding a time dimension.

Further, the statistics of the various parameters of the original features at least include: and the summary value, the mean value, the maximum value, the minimum value, the variation coefficient, the ratio, the standard deviation and the quantile statistic of the original characteristics.

Further, converting the original features according to the client data to derive feature variables with business significance, including:

under different time dimensions, the original features are converted, and feature variables with service significance are derived by classifying and summarizing, calculating the duration or calculating the interruption duration.

Further, in the client data, selecting client data with personal consumption loan amount measurement behavior record in a historical time range as a model training sample, performing machine learning training by using whether a client obtaining the measurement amount subsequently applies for loan as a target variable, and establishing a client response probability prediction model, wherein the client response probability prediction model is used for predicting the probability of applying for loan after the client knows the measurement amount, and the method comprises the following steps:

when machine learning training is carried out, a LightGBM classification model is adopted.

Further, according to the client data and the pre-credit loan amount of the potential client, the probability of applying for loan after the client knows the measuring amount is predicted by using the client response probability prediction model, the potential client is graded according to the probability, and a corresponding marketing strategy is adopted according to the grading result, wherein the marketing strategy comprises the following steps:

according to the probability, dividing the clients in the same probability range into the same grade;

according to the grading result, adopting a first type of marketing strategy for the customers with the probability greater than a first threshold value; and adopting a second type of marketing strategy for the customers with the probability less than the first threshold value.

In a second aspect of an embodiment of the present invention, a marketing response processing system for a potential customer of a personal consumption loan is provided, the system comprising:

the client screening module is used for screening potential clients according to client data and calculating corresponding pre-credit loan amounts in batches;

the response model establishing module is used for selecting client data with personal consumption loan amount measuring and calculating behavior records in a historical time range from the client data as a model training sample, performing machine learning training by taking whether a client obtaining the measured and calculated amount subsequently applies for loan as a target variable, and establishing a client response probability prediction model which is used for predicting the probability of applying for loan after the client knows the measured and calculated amount;

and the marketing response module is used for predicting the probability of applying for loan after the client knows the measuring and calculating limit by using the client response probability prediction model according to the client data and the pre-granted loan limit of the potential client, grading the potential client according to the probability and adopting a corresponding marketing strategy according to the grading result.

In a third aspect of the embodiments of the present invention, a computer device is provided, which includes a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements a marketing response processing method for a potential customer of personal loan consumption when executing the computer program.

In a fourth aspect of the embodiments of the present invention, a computer-readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the computer program implements a marketing response processing method for a potential customer of personal consumption loan.

The marketing response processing method and the marketing response processing system for the personal consumption loan potential customers screen the potential customers according to the customer data, and calculate the corresponding pre-credit loan amount in batches; in the client data, selecting client data recorded with personal consumption loan amount measuring and calculating behaviors in a historical time range as a model training sample, performing machine learning training by taking whether a client obtaining the measured and calculated amount subsequently applies for loan as a target variable, and establishing a client response probability prediction model which is used for predicting the probability of applying for loan after the client knows the measured and calculated amount; according to the client data of the potential client and the pre-credit loan amount, the client response probability prediction model is used for predicting the probability of applying for loan after the client knows the measuring and calculating amount, the potential client is graded according to the probability, and a corresponding marketing strategy is adopted according to the grading result, so that the basis which is easier to understand and operate is provided for the business development of marketers, the differentiated and accurate marketing is realized, and the marketing efficiency and the conversion rate are effectively improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a flow chart of a marketing response processing method of a potential customer for personal loan consumption according to an embodiment of the invention.

Fig. 2 is a flow diagram illustrating screening of potential customers of a personal loan, in accordance with an embodiment of the invention.

Fig. 3 is a schematic diagram of a marketing response processing system architecture for a potential customer of personal loan consumption according to an embodiment of the invention.

FIG. 4 is a block diagram of a client screening module according to an embodiment of the present invention.

Fig. 5 is a schematic diagram of a marketing response processing system of a potential customer for personal loan consumption according to another embodiment of the invention.

Fig. 6 is a schematic structural diagram of a computer device according to an embodiment of the present invention.

Detailed Description

The principles and spirit of the present invention will be described with reference to a number of exemplary embodiments. It is understood that these embodiments are given solely for the purpose of enabling those skilled in the art to better understand and to practice the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

As will be appreciated by one skilled in the art, embodiments of the present invention may be embodied as a system, apparatus, device, method, or computer program product. Accordingly, the present disclosure may be embodied in the form of: entirely hardware, entirely software (including firmware, resident software, micro-code, etc.), or a combination of hardware and software.

In the embodiments of the present invention, terms to be described include:

AUM: AUM (asset Undermanager management), the asset management Scale, is a marker that measures how much a customer contributes to a credit agency. The AUM comprises personal financial assets such as deposit of a client in a financial institution and various investment products purchased by the financial institution, wherein the investment products mainly comprise funds, national debt, insurance, investment financing products issued by the financial institution and the like.

IV: IV (information value), which is a measure of the single predictive ability of a variable to a target event.

LightGBM: compared with other integrated models, the advanced integrated learning model framework has the characteristics of directly supporting class characteristics, optimizing multithreading, supporting high-efficiency parallelism and the like. The core of the LightGBM is an integrated learning model using a Decision tree as a base classifier, namely, a nonlinear model gbdt (gradient Boosting Decision tree) algorithm. The method aims to gradually improve the base learners through an iterative process of a plurality of decision trees until the number of the base learners reaches a target value. Meanwhile, in order to solve the problems of low efficiency and low expansibility of the GBDT, a unilateral Gradient Sampling GOSS (Gradient-based One-Side Sampling) algorithm and an exclusive Feature binding EFB (exclusive Feature bundling) algorithm are adopted, wherein the former excludes most samples with small gradients from the aspect of reducing samples, and the latter binds exclusive features from the aspect of reducing features, combines the two and manages together, so that high-efficiency implementation is ensured.

According to the embodiment of the invention, a marketing response processing method and a marketing response processing system for potential customers of personal loan consumption are provided; the invention mainly generates derived variables with business meaning in batches by a machine learning method according to information of a client contribution degree (AUM), a proxy wage, a credit card, an in-line loan, a personal consumption loan amount measurement and calculation and an application behavior of a client. Performing a correlation test by coverage rate test, calculating a group stability indicator (PSI) and an Information Value (IV) and performing primary screening on the plurality of variables; in a historical time range, selecting client data recorded with personal consumption loan amount measuring and calculating behaviors as a model training sample, and establishing a client response probability prediction model by adopting a LightGBM classification model to predict the probability of applying for loan after the client knows the measuring and calculating amount by taking whether the client who obtains the measuring and calculating amount subsequently applies for loan as a target variable; and converting the probability predicted by the response model into a score to grade the client.

Based on the trained model, the method is applied to potential customers, and differential marketing strategies are adopted according to the grades of the potential customers, so that accurate marketing is realized. The potential customers are selected from the line stock customers which do not deal with the line consumption loan product business at present according to the basic admission rules of the line personal consumption loan; and simultaneously, calculating the pre-credit loan amount of the potential customer in batch according to the amount calculation rule.

The principles and spirit of the present invention are explained in detail below with reference to several representative embodiments of the invention.

Fig. 1 is a flow chart of a marketing response processing method of a potential customer for personal loan consumption according to an embodiment of the invention. As shown in fig. 1, the method includes:

step S1, screening potential customers according to the customer data, and calculating corresponding pre-credit loan amount in batch;

step S2, selecting client data with personal consumption loan amount measuring and calculating behavior record in a historical time range as a model training sample in the client data, performing machine learning training by taking whether a client obtaining the measured and calculated amount subsequently applies for loan as a target variable, and establishing a client response probability prediction model which is used for predicting the probability of applying for loan after the client knows the measured and calculated amount;

and step S3, according to the client data of the potential client and the pre-granted loan amount, predicting the probability of applying for loan after the client knows the measuring amount by using the client response probability prediction model, grading the potential client according to the probability, and adopting a corresponding marketing strategy according to the grading result.

It should be noted that although the operations of the method of the present invention have been described in the above embodiments and the accompanying drawings in a particular order, this does not require or imply that these operations must be performed in this particular order, or that all of the operations shown must be performed, to achieve the desired results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions.

For a more clear explanation of the marketing response processing method of the potential customers for personal loan consumption, the following description is made in conjunction with each step.

And step S1, screening potential customers according to the customer data, and calculating corresponding pre-credit loan amount in batches.

Wherein the customer data comprises at least: client contribution degree, proxy wage information, credit card information, inline loan information, personal consumption loan amount measuring and calculating information and application behavior information of a client; these features may enable the model to learn the client's intent with respect to the individual consuming the loan product.

Specifically, the client contribution (AUM information) of the client includes: AUM source product types (deposit, insurance, financing, etc.), client monthly and daily average AUM values, etc.

The accrual payroll information includes: a sponsor, a transaction amount, a transaction number, etc.

The credit card information includes: credit card available amount, credit card overdue amount, credit card non-overdue balance, etc.

The loan information comprises the current period number of the loan, the date of repayment, the principal of owing, the settlement date of the current period, and the like.

The loan application behavior information comprises click quota checking time, application time and the like.

In this embodiment, the customer data may include derived features in addition to the original features. The specific process of characteristic derivation is as follows:

step S101, generating derived variables related to the business in batches according to client data;

and S102, carrying out preliminary screening on the derived variables through coverage rate inspection and population stability index and information value and correlation inspection.

In step S101, feature variables with business significance may be derived through aggregation and conversion of original features.

When aggregation is adopted for feature derivation, different time windows are divided for running data or detail type data, statistical variables in each window are calculated, and new features are derived; and for the data of the typing variables, calculating the occurrence times of the types and the types of the occurrence types, and deriving new characteristics.

Specifically, statistics such as a summary value, a mean value, a maximum value, a minimum value, a variation coefficient, an occupation ratio, a standard deviation, a quantile and the like of each original feature can be calculated, and a time dimension (about 1/3/6/12 months and the like) is added to generate new features, such as the total amount of deposit type AUM in about 1/3/6 months, the maximum value of unexpired balance of the credit card in about 90/180 days and the like.

The conversion is to perform operations such as sorting and summarizing, storage duration calculation, interruption duration calculation and the like on each original feature to generate new features under different time dimensions, for example, the number of times of checking the sum of the last time is about 1/3/7/30 days, the number of months is continuously increased by the AUM in the past 3/6/12 months, and the like.

The derived variables can be used as training samples of the model in subsequent steps.

Specifically, referring to table 1, when constructing the customer response probability prediction model, a total of 145 derived variables of each type may be used.

TABLE 1 in-mold variables

Referring to fig. 2, a flow chart of screening potential customers for personal loans according to an embodiment of the invention is shown. As shown in fig. 2, the specific process is as follows:

step S201, determining the range of the individual consumption loan products according to the service requirements, the application scenes and the availability of data.

Financial institutions generally have a plurality of individual consumer loan products, and before a response model is established, the product range of the model needs to be determined by combining business requirements, application scenes, data availability and the like.

Step S202, according to the range of personal consumption loan products, a first customer group which accords with the conditions of public admission rules is screened from the customer data; wherein the common admission rule conditions comprise at least: non-blacklist customers, non-grey list customers, anti-money laundering verification pass, no loan overdue exists, non-internal control list customers and are in line with age restrictions.

The first condition to become a potential customer is public admission passage, and public admission rules are important lines of defense for screening high-risk customers.

Step S203, according to the customer data of the first customer group, removing the customers with inventory unrefined personal consumption loan accounts to obtain a second customer group;

the customers with the unbundled personal consumption loan accounts are not suitable for applying for loans, otherwise, the risk of default is increased, so that the customers are non-potential customers, and the customers need to be removed before the amount is calculated.

And step S204, carrying out pre-granted credit limit measurement and calculation according to the client data of the second client group, and reserving the client with the measurement and calculation result larger than 0 as a potential client.

When the amount is measured and calculated, the main data sources are the client contribution degree, the proxy wage information and the house property information of the second client group; and carrying out measurement and calculation of the pre-granted credit limit based on the data.

Through public admission, no inventory outstanding personal consumption loan accounts exist, and the customers with the final quota larger than 0 are potential customers and are samples for subsequently establishing a response model.

Step S2, a response model is established.

In the client data, selecting client data recorded with personal consumption loan amount measuring and calculating behaviors in a historical time range as a model training sample, performing machine learning training by taking whether a client obtaining the measured and calculated amount subsequently applies for loan as a target variable, and establishing a client response probability prediction model, wherein the client response probability prediction model is used for predicting the probability of applying for loan after the client knows the measured and calculated amount.

In this embodiment, a LightGBM classification model is used in the machine learning training. The LightGBM algorithm is an integrated algorithm, is improved on the basis of the traditional GBDT and is an efficient implementation of the GBDT, and the main idea of the algorithm is to use iterative training of a decision tree to obtain an optimal model, use a negative gradient of a loss function as a residual error approximate value of the current decision tree and fit a new decision tree. LightGBM replaces the Pre-ordering method (Pre-sorted) in the GBDT algorithm with the histogram algorithm, occupies lower memory and has lower complexity of data separation. The model has the advantages of good training effect, difficulty in overfitting, high speed and the like.

Furthermore, after a customer response probability prediction model is established, some indexes of the model can be calculated, the model is evaluated, and the optimal model is selected; the specific process is as follows:

setting a test set according to the customer data;

applying the established customer response probability prediction model to the test set, and calculating the evaluation index of the customer response probability prediction model to obtain the root mean square error and the fitting degree of the model;

wherein, the Root Mean Square Error (RMSE) is used to measure the deviation between the observed value and the true value, reflecting the precision of measurement.

R2 score (the coeffient of determination), judged is the degree of fit of the prediction model to the real data.

And evaluating the model according to the root mean square error and the fitting degree of the model, and selecting a customer response probability prediction model. Wherein, the smaller RMSE represents the better model fitting effect, and the larger R2 Score represents the better model fitting effect. And determining a stable model with good fitting effect according to the two evaluated indexes by applying the model obtained by training the training sample to a test set.

In this embodiment, the obtained probability is converted into a score, and the customer is graded according to the score, wherein the application probability levels of different gears are different.

And combining the promotion degree of each fractional interval section, the customer can be classified into A, B, C, D, E, F sixth gear, the response probability of the customer in the A gear is the highest, and the response probabilities of other gears are reduced in sequence.

As shown in Table 2, the fractional block segments of A, B, C, D, E, F sixth gear are (700,850], (650,700), (600,650), (550, 600), (500, 550) and (299,500), respectively.

TABLE 2 customer grading Table

The specific grading manner, the score interval corresponding to the grading, and the score corresponding to the predicted response probability may be adjusted according to actual needs, and this is only an exemplary example.

The response model scores and the grades of the potential customers can be provided to business personnel as data items, and the business personnel can implement differential marketing strategies for the customers with different application probability levels.

For example, the method carries out telemarketing to the customers with higher application probability, and carries out short message and mail marketing to the customers with lower application probability, thereby improving the marketing efficiency and the conversion rate and realizing accurate marketing.

Specifically, for the A, B, C graded customers, since the response probability is above 50%, the salesperson can carry out marketing of personal consumption loan and related businesses by a telephone marketing mode with good communication effect.

And for D, E, F graded customers, marketing can be carried out by means of short messages, mails and the like.

Compared with the prior art, the marketing response processing method for the potential customers of the personal consumption loan has at least the following advantages:

by establishing a customer response probability prediction model, the probability of applying for loan after the customer knows the measuring and calculating amount is predicted, and the probability of applying for loan after the customer knows the pre-granted credit amount through marketing activities is evaluated.

According to the probability predicted by the response model, the score is converted into the score to grade the client, so that the basis which is easier to understand and operate is provided for the adjustment of marketing strategies and the business development of marketing personnel, the differential and accurate marketing is carried out, and the marketing conversion rate is effectively improved.

By adopting the LightGBM classification model, the algorithm has good training effect, is not easy to over-fit, has high speed and effectively improves the overall efficiency.

Having described the method of an exemplary embodiment of the present invention, a marketing response processing system for a person consuming loan potential customer of an exemplary embodiment of the present invention is next described with reference to fig. 3 through 5.

The implementation of the marketing response processing system for the potential customers for personal loan consumption can be referred to the implementation of the method, and repeated details are omitted. The term "module" or "unit" used hereinafter may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.

Based on the same inventive concept, the invention also provides a marketing response processing system for potential customers of personal loan consumption, as shown in fig. 3, the system comprises:

the client screening module 310 is used for screening potential clients according to the client data and calculating corresponding pre-credit loan amount in batches;

the response model establishing module 320 is used for selecting client data with personal consumption loan amount measuring and calculating behavior record in a historical time range from the client data as a model training sample, performing machine learning training by taking whether a client obtaining the measured and calculated amount subsequently applies for loan as a target variable, and establishing a client response probability prediction model which is used for predicting the probability of applying for loan after the client knows the measured and calculated amount;

wherein, when the machine learning training is carried out, a LightGBM classification model is adopted. The LightGBM algorithm is an integrated algorithm, is improved on the basis of the traditional GBDT and is an efficient implementation of the GBDT, and the main idea of the algorithm is to use iterative training of a decision tree to obtain an optimal model, use a negative gradient of a loss function as a residual error approximate value of the current decision tree and fit a new decision tree. LightGBM replaces the Pre-ordering method (Pre-sorted) in the GBDT algorithm with the histogram algorithm, occupies lower memory and has lower complexity of data separation. The model has the advantages of good training effect, difficulty in overfitting, high speed and the like.

And the marketing response module 330 is used for predicting the probability of applying for loan after the client knows the measuring and calculating amount by using the client response probability prediction model according to the client data of the potential client and the pre-granted loan amount, grading the potential client according to the probability, and adopting a corresponding marketing strategy according to the grading result.

It should be noted that although several modules of the marketing response processing system for a person consuming a loan potential customer are mentioned in the above detailed description, such partitioning is merely exemplary and not mandatory. Indeed, the features and functionality of two or more of the modules described above may be embodied in one module according to embodiments of the invention. Conversely, the features and functions of one module described above may be further divided into embodiments by a plurality of modules.

In one embodiment, the customer data includes at least: client contribution degree, proxy wage information, credit card information, inline loan information, personal consumption loan amount measurement and calculation information and application behavior information of the client.

In an embodiment, referring to fig. 4, a schematic diagram of an architecture of a client screening module according to an embodiment of the present invention is shown. As shown in fig. 4, the customer filtering module 310 includes:

a product range determining unit 311, configured to determine a range of the personal loan product to be consumed according to the service requirement, the application scenario, and the availability of data;

an admission screening unit 312, configured to screen, according to the range of the personal consumption loan products, a first customer group that meets the conditions of the common admission rules from the customer data;

wherein the common admission rule conditions comprise at least: non-blacklist customers, non-grey list customers, anti-money laundering verification pass, no loan overdue exists, non-internal control list customers and are in line with age restrictions.

The client removing unit 313 is used for removing clients with stored amounts of unrefined personal consumption loan accounts according to the client data of the first client group to obtain a second client group;

and the amount measuring and calculating unit 314 is used for measuring and calculating the pre-granted credit amount according to the client data of the second client group, and reserving the client with the measuring and calculating result larger than 0 as a potential client.

In an embodiment, the credit calculation unit 314 is specifically configured to: and pre-granting credit line budget according to the client contribution degree, the proxy wage information and the house property information of the second client group.

In another embodiment, referring to fig. 5, a marketing response processing system architecture of a potential customer for loan consumption by an individual in accordance with another embodiment of the invention is shown. As shown in fig. 5, the system further includes a derived variable module 340, specifically configured to:

The aggregation and conversion of the original features can be performed according to the client data, and feature variables with business significance are derived.

In one embodiment, marketing response module 330 is specifically configured to:

according to the client data of potential clients and the pre-credit loan amount, predicting the probability of applying for loan of the clients after knowing the measuring and calculating amount by using the client response probability prediction model;

Further, after the response model establishing module 320 establishes the customer response probability prediction model, some indexes of the model can be calculated, the model is evaluated, and the optimal model is selected; the specific process is as follows:

setting a test set according to the customer data;

applying the established customer response probability prediction model to the test set, and calculating an evaluation index of the customer response probability prediction model to obtain a Root Mean Square Error (RMSE) and a fitting degree (R2 Score) of the model;

Based on the aforementioned inventive concept, as shown in fig. 6, the present invention further provides a computer device 600, which comprises a memory 610, a processor 620 and a computer program 630 stored in the memory 610 and operable on the processor 620, wherein the processor 620 executes the computer program 630 to implement the aforementioned marketing response processing method for potential customers of personal loan consumption.

Based on the foregoing inventive concept, the present invention provides a computer-readable storage medium storing a computer program, which when executed by a processor implements the aforementioned marketing response processing method for potential customers of personal loan consumption.

In the marketing response processing process of the potential client for the personal consumption loan, whether the client for measuring and calculating the quota subsequently applies for the loan or not is taken as a target variable; besides the common characteristic dimensions of the client contribution degree (AUM), the substituted payroll, the credit card and the in-line loan, the characteristics related to the calculation of the personal consumption loan amount and the application behavior are added in the characteristics, so that the model learns the intention of the client about the personal consumption loan products. Specifically, a LightGBM classification model can be adopted to establish a customer response probability prediction model so as to predict the probability of applying for loan after the customer knows the measurement amount. And converting the probability predicted by the response model into a score to grade the client. And subsequently, the trained model is applied to potential customers, and a differentiated marketing strategy is adopted according to the grading of the potential customers, so that accurate marketing is realized.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A marketing response processing method for a potential customer of a personal consumption loan, the method comprising:

2. The method of claim 1, wherein the customer data includes at least: client contribution degree, proxy wage information, credit card information, inline loan information, personal consumption loan amount measurement and calculation information and application behavior information of the client.

3. The marketing response processing method for potential customers of personal loan consumption of claim 1, wherein the step of screening potential customers according to the customer data and calculating corresponding pre-granted loan amounts in batches comprises:

4. The method of claim 3, wherein the common admission rule conditions include at least: non-blacklist customers, non-grey list customers, anti-money laundering verification pass, no loan overdue exists, non-internal control list customers and are in line with age restrictions.

5. The marketing response processing method for the potential customers for personal loan consumption of claim 3, wherein the pre-granted credit limit calculation is performed according to the customer data of the second customer group, and the customers with the calculation result larger than 0 are reserved as the potential customers, comprising:

6. The method of claim 2, further comprising:

7. The method of claim 6, wherein the step of batch generating business related derivative variables from customer data comprises:

8. The method of claim 7, wherein the step of deriving the characteristic variables with business significance by aggregating the raw characteristics according to the client data comprises:

9. The method of claim 7, wherein the statistics of the plurality of parameters of the raw characteristics include at least: and the summary value, the mean value, the maximum value, the minimum value, the variation coefficient, the ratio, the standard deviation and the quantile statistic of the original characteristics.

10. The method of claim 7, wherein the deriving the characteristic variables with business significance by performing a conversion of the original characteristics according to the client data comprises:

11. The marketing response processing method of potential customers for personal consumption loan as claimed in claim 1, wherein the customer data is selected as a model training sample in a historical time range, and the model training is performed to determine whether the customers who obtain the measurement limit have continued loan application as a target variable, so as to establish a customer response probability prediction model, and the customer response probability prediction model is used for predicting the probability of applying loan after the customers know the measurement limit, and the marketing response processing method comprises the following steps:

12. The marketing response processing method of the potential customer for the personal consumption loan of claim 1, wherein the probability of applying for the loan of the customer after knowing the measurement limit is predicted by using the customer response probability prediction model according to the customer data and the pre-granted loan limit of the potential customer, the potential customer is graded according to the probability, and a corresponding marketing strategy is adopted according to the grading result, comprising the following steps:

13. A marketing response processing system for a potential customer of a personal consumption loan, the system comprising:

14. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any one of claims 1 to 12 when executing the computer program.

15. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, implements the method of any one of claims 1 to 12.