Credit card customer value layering method and device based on RFM model

Abstract

The invention discloses a credit card customer value layering method and device based on an RFM model in the technical field of data analysis. The credit card customer value layering method comprises the steps of obtaining credit card transaction information of a customer in a plurality of deadlines; and performing multi-dimensional classification processing on the credit card transaction information to obtain index data. According to the method, the index data are obtained through classification processing after the credit card transaction information of the customer is obtained, the determined dimension evaluation data are obtained according to comparison of the index data and the index threshold value, the value hierarchical data of the customer is determined after the dimension evaluation data are obtained, evaluation data of more dimensions can be obtained, and the user experience is improved. It is guaranteed that the obtained value layering data granularity is refined enough, the evaluation dimension is more comprehensive, then more accurate value customers are obtained, auxiliary reference information is provided for bank operation resource input, classification and distinguished operation are conducted for the customers, and bank resources can be optimally configured.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and specifically to a method and device for stratifying the value of credit card customers based on the RFM model. Background Art

[0002] A credit card is a credit certificate issued by a commercial bank or a credit card company, mainly used to prove the credit status of the cardholder. For the credit card users of a bank, the value difference is judged through the RFM model. Among them, the RFM model is an important tool and means for measuring customer value, and it classifies customers through three behavioral indicators of customers.

[0003] At present, the stratification of the value of credit card customers has insufficient refinement of the evaluation dimension granularity for bank credit card users, which is inconvenient for accurately positioning the value stratification of credit card users and is not conducive to providing auxiliary reference information for the investment of bank operation resources, and the use effect is not good. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and device for stratifying the value of credit card customers based on the RFM model to solve the problem that the refinement degree of the value stratification dimension of credit card users is insufficient and it is inconvenient to accurately position the value stratification of credit card users as mentioned above.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] In the first aspect, the present invention provides a method for stratifying the value of credit card customers based on the RFM model, including: obtaining the credit card transaction information of a customer within a number of periods; performing multi-dimensional classification processing on the credit card transaction information to obtain index data; comparing the index data with an index threshold to determine the dimension evaluation data corresponding to the customer; determining the value stratification data according to the dimension evaluation data; obtaining the historical risk data of the credit card of the customer within a number of periods, excluding the data in the value stratification data whose historical risk data exceeds the preset risk threshold, and obtaining risk analysis data, processing and outputting the risk analysis data to identify credit card value customers.

[0007] As a further solution of the present invention: the obtaining of the credit card transaction information of a customer within a number of periods includes; the credit card transaction information includes credit card consumption amount data, credit card consumption frequency data, and the data of the last transaction date of the credit card; the credit card consumption amount data is the actual receivable amount data, and the credit card consumption amount data does not include the swipe fee; wherein the time span of the number of periods is greater than or equal to a six-month period.

[0008] As a further solution of the present invention: classifying the credit card transaction information in multiple dimensions to obtain index data, including: classifying the credit card transaction amount into monthly average expenditure amount data and the ratio of installment amount to cash withdrawal amount data to obtain index data;

[0009] Classifying the credit card consumption frequency data into monthly average utilization times data and the ratio of utilization months to several periods data to obtain index data; comparing the credit card's last trading day data with the current date data to obtain duration data, and the duration data is index data, where the duration data is the current date data minus the credit card's last trading day data.

[0010] As a further solution of the present invention: comparing the index data with index thresholds to determine the dimension evaluation data corresponding to the customer, including: the index thresholds include the average expenditure amount threshold, the ratio of installment amount to cash withdrawal amount threshold, the monthly average utilization times threshold, the ratio of utilization months to several periods threshold, and the duration threshold; the dimension evaluation data corresponding to the customer includes consumption amount dimension data, consumption frequency dimension data, and utilization duration dimension data. The consumption amount dimension data and the consumption frequency dimension data are both in a high or low state, and the utilization duration dimension data is in an active and dormant state; comparing the average expenditure amount data with the average expenditure amount threshold, and comparing the ratio of installment amount to cash withdrawal amount data with the ratio of installment amount to cash withdrawal amount threshold; if the average expenditure amount data is less than or equal to the average expenditure amount threshold and the ratio of installment amount to cash withdrawal amount data is less than or equal to the ratio of installment amount to cash withdrawal amount threshold, then determine that the consumption amount dimension data is low, otherwise determine that the consumption amount dimension data is high; comparing the monthly average utilization times data with the monthly average utilization times threshold, and comparing the ratio of utilization months to several periods data with the ratio of utilization months to several periods threshold; if the monthly average utilization times data is less than or equal to the monthly average utilization times threshold and the ratio of utilization months to several periods data is less than the ratio of utilization months to several periods threshold, then determine that the consumption frequency dimension data is low, otherwise determine that the consumption frequency dimension data is high; comparing the duration data with the duration threshold, if the duration data is greater than the duration threshold, then determine that the utilization duration dimension data is dormant, otherwise it is active.

[0011] As a further solution of the present invention: the priority of comparing the average expenditure amount data is higher than the priority of comparing the ratio of installment amount to cash withdrawal amount data; the priority of comparing the monthly average utilization times data is higher than the priority of comparing the ratio of utilization months to several periods data.

[0012] As a further solution of the present invention: determining value stratification data according to the dimension evaluation data includes; the value stratification data includes important value customers, important development customers, important retention customers, important attention customers and ordinary customers; if the data of the utilization duration dimension is active and the data of both the consumption amount dimension and the consumption frequency dimension are high, it is determined as an important value customer; if the data of the utilization duration dimension is active and one of the data of the consumption amount dimension and the consumption frequency dimension is high, it is determined as an important development customer; if the data of the utilization duration dimension is dormant and at least one of the data of the consumption amount dimension and the consumption frequency dimension is high, it is determined as an important retention customer; if the data of both the consumption amount dimension and the consumption frequency dimension are low and the data of the utilization duration dimension is active, it is determined as an important attention customer; if the data of both the consumption amount dimension and the consumption frequency dimension are low and the data of the utilization duration dimension is dormant, it is determined as an ordinary customer.

[0013] As a further solution of the present invention: the important development customers include the first important development customer and the second important development customer; when the data of the consumption amount dimension is high, it is determined as the first important development customer; when the data of the consumption frequency dimension is high, it is determined as the second important development customer.

[0014] As a further solution of the present invention: the important retention customers include the first important retention customer, the second important retention customer and the third important retention customer; when the data of both the consumption amount dimension and the consumption frequency dimension are high, it is determined as the first important retention customer; when the data of the consumption amount dimension is high and the data of the consumption frequency dimension is low, it is determined as the second important retention customer; when the data of the consumption amount dimension is low and the data of the consumption frequency dimension is high, it is determined as the third important retention customer.

[0015] As a further solution of the present invention: obtaining the historical risk data of the customer's credit card within a number of periods, excluding the data in the value stratification data whose historical risk data exceeds the preset risk threshold, and obtaining risk analysis data, processing and outputting the risk analysis data, including; the historical risk data includes the historical overdue days data of the credit card, and the risk threshold includes the overdue days threshold; excluding the data in the value stratification data whose historical overdue days data is greater than or equal to the overdue days threshold, and obtaining risk analysis data; processing the risk analysis data to obtain the excluded value stratification data, the corresponding customer ratio data in the value stratification data and the corresponding income ratio data in the value stratification data; outputting the excluded value stratification data, the customer ratio data and the income ratio data to identify credit card value customers.

[0016] In a second aspect, the present invention provides a credit card customer value stratification device based on the RFM model, including an acquisition module, a multi-dimensional classification processing module, a comparison module, a value stratification module, and a risk output module;

[0017] The acquisition module is used to acquire the credit card transaction information of customers within a certain period; the multi-dimensional classification processing module performs multi-dimensional classification processing on the credit card transaction information to obtain index data; the comparison module compares the index data with an index threshold to determine the dimension evaluation data corresponding to the customer; the value stratification module determines the value stratification data according to the dimension evaluation data;

[0018] The risk output module acquires the historical risk data of the credit card of the customer within a certain period, excludes the data in the value stratification data whose historical risk data exceeds the preset risk threshold, and obtains risk analysis data, processes and outputs the risk analysis data to identify credit card value customers.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0020] 1. In the present invention, after acquiring the credit card transaction information of the customer, index data is obtained through classification processing. Based on the comparison of the index data and the index threshold, the determined dimension evaluation data is obtained. After obtaining the dimension evaluation data, the value stratification data of the customer is determined. This method can obtain more-dimensional evaluation data, ensure that the obtained value stratification data has a sufficiently refined granularity, and the evaluation dimension is more comprehensive. Furthermore, more accurate value customers can be obtained, providing auxiliary reference information for the bank's operation resource investment, classifying customers and differentiating operations, and enabling the optimal allocation of bank resources.

[0021] 2. In the present invention, when determining the dimension evaluation data of the customer, the average amount of money used is used as an important indicator to measure the customer's consumption ability and credit level, and is given the highest priority. During the process of evaluating the customer value stratification, the average amount of money used data is preferentially determined. By comparing it with the preset threshold, the customer's consumption ability and credit level are initially judged. Similarly, the monthly average number of times of use data, as an important indicator to measure the customer's activity, has a higher priority in the comparison result than the ratio of the number of months of use to the data of a certain period. This method reduces the overall calculation amount, can quickly determine the dimension evaluation data, and has good use effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a schematic flow chart of the method of the present invention;

[0023] Figure 2 is a schematic connection diagram of the device of the present invention.

[0024] In the figure: 1. Acquisition module; 2. Multidimensional classification processing module; 3. Comparison module; 4. Value stratification module; 5. Risk output module. Specific implementation manner

[0025] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0026] Embodiment:

[0027] Please refer to Figure 1 , in the embodiment of the present invention, a method for stratifying the value of credit card customers based on the RFM model includes:

[0028] S1: Obtain the credit card transaction information of the customer within a certain period.

[0029] S2: Perform multidimensional classification processing on the credit card transaction information to obtain index data.

[0030] S3: Compare the index data with the index threshold to determine the dimension evaluation data corresponding to the customer.

[0031] S4: Determine the value stratification data according to the dimension evaluation data.

[0032] S5: Obtain the historical risk data of the customer's credit card within a certain period, exclude the data in the value stratification data whose historical risk data exceeds the preset risk threshold, and obtain risk analysis data. Process and output the risk analysis data to identify credit card value customers.

[0033] In the present invention, after obtaining the credit card transaction information of customers, it is classified and processed to obtain index data. Based on the comparison between the index data and the index thresholds, the determined dimension evaluation data is obtained. After obtaining the dimension evaluation data, the value stratification data of the customers is determined. This method can obtain more dimensional evaluation data, ensure that the obtained value stratification data has a sufficiently refined granularity, and the evaluation dimensions are more comprehensive. Furthermore, more accurate valuable customers can be obtained, providing auxiliary reference information for the investment of bank operation resources. Classifying and differentiating operations for customers can optimize the allocation of bank resources, with good use effects. And after obtaining the value stratification data of credit card users, after finely dividing the value stratification, according to the historical risk data of the credit card within the corresponding period of the customers, the corresponding historical risk data is compared with the preset risk threshold. When the value of the historical risk data exceeds the risk threshold, the value stratification data exceeding the risk threshold is excluded. Furthermore, the credit card customers exceeding the preset risk are excluded, reducing the overall risk of the bank, with good use effects.

[0034] Preferably, obtain the credit card transaction information of customers within several periods, including;

[0035] The credit card transaction information includes credit card consumption amount data, credit card consumption frequency data, and the date of the last credit card transaction;

[0036] The credit card consumption amount data is the actual receivable amount data, and the credit card consumption amount data does not include the swipe fee;

[0037] Among them, the time span of several periods is greater than or equal to six months.

[0038] Specifically, the credit card consumption frequency data records the number of transactions of customers within a certain period of time, reflecting the activity of customers; the date of the last credit card transaction data records the date when customers last used the credit card for transactions, which helps the bank understand the latest transaction dynamics of customers;

[0039] Through the comprehensive analysis of these data, the bank can more comprehensively understand the consumption habits, credit status, and potential risks of customers, providing strong support for subsequent value stratification and risk assessment;

[0040] Furthermore, selecting a time span of greater than or equal to six months for data analysis can ensure the stability and reliability of the data, avoiding misjudgment caused by short-term fluctuations. Among them, several periods can be eight months, ten months, or twelve months, and are selected according to the actual credit card transaction time span.

[0041] Preferably, perform multi-dimensional classification processing on the credit card transaction information to obtain index data, including;

[0042] Classify the credit card transaction limit into the monthly average expenditure amount data and the installment amount ratio to cash withdrawal amount data to obtain the index data;

[0043] Classify the credit card consumption frequency data into the monthly average usage times data and the usage month ratio to several-term data to obtain the index data;

[0044] Compare and process the credit card's last trading date data and the current date data to obtain the duration data, and the duration data is the index data, where the duration data is the current date data minus the credit card's last trading date data.

[0045] Specifically, the monthly average expenditure amount data reflects the average amount that customers use the credit card to consume each month, which helps the bank evaluate the customers' consumption ability and credit needs; the installment amount ratio to cash withdrawal amount data reveals the proportion of customers who choose installment payment or cash withdrawal when using the credit card, which to a certain extent reflects the customers' capital liquidity and debt repayment pressure;

[0046] The monthly average usage times data records the average number of times that customers use the credit card to conduct transactions each month, further subdividing the customers' activity level; the usage month ratio to several-term data shows the ratio of the months when customers use the credit card within a specific time span, which helps the bank identify the customers' seasonal consumption habits or potential risks;

[0047] The duration data directly reflects the length of time since the customers' last credit card transaction, which is of great significance for the bank to judge the customers' activity level and potential churn risk. Through the comprehensive analysis of these index data, the bank can gain a deeper insight into the customers' consumption behavior, credit status and potential needs, providing a scientific basis for formulating personalized marketing strategies and risk management plans.

[0048] Preferably, compare the index data with the index thresholds to determine the dimension evaluation data corresponding to the customers, including;

[0049] The index thresholds include the average expenditure amount threshold, the installment amount ratio to cash withdrawal amount threshold, the monthly average usage times threshold, the usage month ratio to several-term threshold and the duration threshold;

[0050] The dimension evaluation data corresponding to the customers includes the consumption limit dimension data, the consumption frequency dimension data and the usage duration dimension data. The consumption limit dimension data and the consumption frequency dimension data are both in the high or low state, and the usage duration dimension data is in the active and dormant states;

[0051] Compare the average expenditure amount data with the average expenditure amount threshold, and compare the installment amount ratio to cash withdrawal amount data with the installment amount ratio to cash withdrawal amount threshold;

[0052] If the average disbursement amount data is less than or equal to the average disbursement amount threshold and the installment amount to cash withdrawal amount data is less than or equal to the installment amount to cash withdrawal amount threshold, then determine that the consumption limit dimension data is low; otherwise, determine that the consumption limit dimension data is high.

[0053] Compare the monthly average utilization times data with the monthly average utilization times threshold, and compare the utilization month ratio to several term data with the utilization month ratio to several term threshold.

[0054] If the monthly average utilization times data is less than or equal to the monthly average utilization times threshold comparison and the utilization month ratio to several term data is less than the utilization month ratio to several term threshold, then determine that the consumption frequency dimension data is low; otherwise, determine that the consumption frequency dimension data is high.

[0055] Compare the duration data with the duration threshold. If the duration data is greater than the duration threshold, then determine that the utilization duration dimension data is dormant; otherwise, it is active.

[0056] Specifically, by adding dimensions for comparing the average disbursement amount data with the average disbursement amount threshold, the monthly average utilization times data with the monthly average utilization times threshold, and the duration data with the duration threshold, the accuracy of the value stratification data is increased, providing accurate auxiliary reference information for subsequent bank operation resource investment.

[0057] Furthermore, convert these data into visual charts or reports for quick understanding and analysis. When generating visual charts, different colors or icons can be used to represent different evaluation states. For example, use green to represent high or active states, and use red to represent low or dormant states. At the same time, we can also mark the specific values or percentages of each dimension data in the chart for further analysis and comparison.

[0058] Furthermore, the dimension evaluation data can also include the customer's basic information, historical transaction records, repayment records, etc., so as to comprehensively understand the customer's credit status and risk level, providing a basis for subsequent credit decisions.

[0059] Preferably, the priority of comparing the average disbursement amount data is higher than the priority of comparing the installment amount to cash withdrawal amount data.

[0060] The priority of comparing the monthly average utilization times data is higher than the priority of comparing the utilization month ratio to several term data.

[0061] Specifically, when determining the dimensional evaluation data of customers, the average amount used data, as an important indicator for measuring customers' consumption ability and credit level, is given the highest priority. During the process of evaluating customer value stratification, the average amount used data is preferentially determined. By comparing it with a preset threshold, the consumption ability and credit level of customers are initially judged. This method reduces the overall calculation volume, can quickly determine the dimensional evaluation data, and has a good usage effect.

[0062] The installment amount ratio and cash withdrawal amount data, although they can also reflect customers' consumption behaviors and credit statuses, are compared after the average amount used data because their information volume and accuracy are relatively low. Such a setting helps us more accurately grasp customers' credit statuses and avoid making wrong judgments due to the interference of secondary information.

[0063] Similarly, as an important indicator for measuring customers' activity levels, the monthly average usage frequency data also has a higher priority in the comparison results than the usage month ratio for several periods data. By giving priority to the monthly average usage frequency data, we can more quickly identify active customers and dormant customers, providing a basis for subsequent customer management and marketing strategy formulation.

[0064] Preferably, determining the value stratification data according to the dimensional evaluation data includes;

[0065] The value stratification data includes important value customers, important development customers, important retention customers, important attention customers, and ordinary customers;

[0066] If the usage duration dimension data is active and both the consumption amount dimension data and the consumption frequency dimension data are high, it is determined as an important value customer;

[0067] If the usage duration dimension data is active and one of the consumption amount dimension data and the consumption frequency dimension data is high, it is determined as an important development customer;

[0068] If the usage duration dimension data is dormant and at least one of the consumption amount dimension data and the consumption frequency dimension data is high, it is determined as an important retention customer;

[0069] If both the consumption amount dimension data and the consumption frequency dimension data are low and the usage duration dimension data is active, it is determined as an important attention customer;

[0070] If both the consumption amount dimension data and the consumption frequency dimension data are low and the usage duration dimension data is dormant, it is determined as an ordinary customer.

[0071] Specifically, by using multi-dimensional data to determine the value stratification data corresponding to customers, customers are classified into important value customers, important development customers, important retention customers, important attention customers, and ordinary customers, which facilitates the customer identification in the subsequent output results.

[0072] Preferably, important development customers include first important development customers and second important development customers;

[0073] When the data in the consumption amount dimension is high, it is determined as the first important development customer;

[0074] When the data in the consumption frequency dimension is high, it is determined as the second important development customer.

[0075] Preferably, important retention customers include first important retention customers, second important retention customers, and third important retention customers;

[0076] When both the data in the consumption amount dimension and the data in the consumption frequency dimension are high, it is determined as the first important retention customer;

[0077] When the data in the consumption amount dimension is high and the data in the consumption frequency dimension is low, it is determined as the second important retention customer;

[0078] When the data in the consumption amount dimension is low and the data in the consumption frequency dimension is high, it is determined as the third important retention customer.

[0079] Specifically, the classification of important retention customers is further refined and graded, which helps the bank to more accurately identify the potential value and churn risk of customers. The first important retention customers, as a group with both high consumption ability and high consumption frequency, their churn will have a greater impact on the bank, so they should be the top priority of the retention work. Although the second important retention customers have a high consumption amount, their consumption frequency is relatively low, which may mean that the customer's purchase habits or needs have changed. The bank needs to formulate corresponding retention strategies through in-depth understanding and analysis. For the third important retention customers, although their consumption frequency is high, their consumption amount is low. They may be loyal users of the bank, but due to economic ability or other factors, their consumption ability is limited. For such customers, the bank can consolidate and deepen the relationship with them by providing cost-effective products or services, as well as enhancing customer loyalty programs.

[0080] The specific table is shown as follows:

[0081]

[0082] Preferably, obtain the historical risk data of the customer's credit card within a certain period, exclude the data in the value stratification data whose historical risk data exceeds the preset risk threshold, and obtain the risk analysis data. Process and output the risk analysis data, including;

[0083] The historical risk data includes historical overdue days data regarding credit cards, and the risk threshold includes an overdue days threshold;

[0084] Exclude the data in the value stratification data where the historical overdue days data is greater than or equal to the overdue days threshold, and obtain risk analysis data;

[0085] Process the risk analysis data to obtain the value stratification data after exclusion, the corresponding customer ratio data in the value stratification data, and the corresponding income ratio data in the value stratification data,

[0086] Output the value stratification data after exclusion, the customer ratio data, and the income ratio data to identify credit card valuable customers.

[0087] Specifically, this step further screens out customer groups with low risk and high value. In the credit card system of the bank, the historical overdue days data is an important indicator for measuring customer credit risk. By setting a reasonable overdue days threshold, customers with poor credit records can be excluded, ensuring the accuracy and effectiveness of subsequent analysis;

[0088] Furthermore, the processing of risk analysis data not only considers the overdue days, but can also combine other risk factors, such as the customer's repayment behavior, credit limit usage, etc., to comprehensively evaluate the customer's credit risk.

[0089] After excluding high-risk customers, stratifying the remaining customers by value can more accurately locate customers at different value levels and provide them with customized services. The output of the customer ratio data and the income ratio data helps the bank understand the proportion of customers at each value level and their contribution to the total income, providing strong support for the bank's resource allocation and marketing strategy formulation.

[0090] In this way, the bank can not only identify high-value credit card customers, but also effectively reduce credit risk, improve the overall customer quality and profitability.

[0091] Please refer to Figure 2 , a credit card customer value stratification device based on the RFM model, including: an acquisition module 1, a multi-dimensional classification processing module 2, a comparison module 3, a value stratification module 4, and a risk output module 5;

[0092] The acquisition module 1, the acquisition module 1 is used to acquire the credit card transaction information of customers within a number of periods;

[0093] The multi-dimensional classification processing module 2, the multi-dimensional classification processes the credit card transaction information to obtain index data;

[0094] Comparison module 3 compares the metric data with the metric threshold to determine the dimension evaluation data corresponding to the customer;

[0095] Value stratification module 4 determines the value stratification data according to the dimension evaluation data;

[0096] Risk output module 5 obtains the historical risk data of the customer's credit card within a number of periods, excludes the data in the value stratification data where the historical risk data exceeds the preset risk threshold, and obtains the risk analysis data. The risk analysis data is processed and output to identify the credit card value customers.

[0097] Specifically, after the acquisition module 1 acquires the credit card transaction information of the customer, and the multi-dimensional classification processing module 2 classifies and processes it to obtain the metric data, the comparison module 3 compares according to the metric data and the metric threshold, and then obtains the determined dimension evaluation data. After obtaining the dimension evaluation data, the value stratification module 4 determines the value stratification data of the customer to obtain more dimension evaluation data, ensuring that the obtained value stratification data has a sufficiently refined granularity and a more comprehensive evaluation dimension, and then obtaining more accurate value customers, providing auxiliary reference information for the bank's operation resource investment, classifying and differentiating the operation for the customers, which can optimize the allocation of bank resources and has a good use effect. And after obtaining the value stratification data of the credit card users, after finely dividing the value stratification, the risk output module 5 compares the corresponding historical risk data with the preset risk threshold according to the historical risk data of the customer's credit card within the corresponding period. When the value of the historical risk data exceeds the risk threshold, the value stratification data exceeding the risk threshold is excluded, and then the credit card customers exceeding the preset risk are excluded, reducing the overall risk of the bank and having a good use effect.

[0098] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes an equivalent substitution or change, and should be covered by the protection scope of the present invention.

Claims

1. A credit card customer value stratification method based on the RFM model, characterized in that: include: Obtain credit card transaction information of customers within a certain period; Perform multi-dimensional classification processing on the credit card transaction information to obtain index data; Compare the indicator data with the indicator threshold to determine the dimension evaluation data corresponding to the customer; Determine value stratification data based on the dimension evaluation data; Obtain historical risk data of the customer's credit card within a certain period of time, exclude the historical risk data in the value stratification data that exceeds the preset risk threshold, and obtain risk analysis data, process and output the risk analysis data to identify credit card value customers.

2. The credit card customer value stratification method based on the RFM model according to claim 1 is characterized in that: The obtaining of credit card transaction information of customers within a certain period of time includes: The credit card transaction information includes credit card spending limit data, credit card spending frequency data and credit card last transaction date data; The credit card spending limit data is the actual amount receivable data, and the credit card spending limit data does not include the card swiping fee; The time span of the several periods mentioned is greater than or equal to six months.

3. The credit card customer value stratification method based on the RFM model according to claim 2 is characterized in that: The multi-dimensional classification processing of the credit card transaction information to obtain index data includes: Classify the credit card transaction amount into monthly average expenditure amount data and installment amount to cash withdrawal amount data to obtain index data; Classify the credit card consumption frequency data into monthly average usage data and usage month to several term data to obtain index data; The last transaction date data of the credit card is compared with the current date data to obtain duration data, which is indicator data, wherein the duration data is the current date data minus the last transaction date data of the credit card.

4. The credit card customer value stratification method based on the RFM model according to claim 3 is characterized in that: The comparing the indicator data with the indicator threshold to determine the dimension evaluation data corresponding to the customer includes: The index thresholds include the average expenditure amount threshold, the installment amount to cash withdrawal amount threshold, the average monthly mobilization times threshold, the mobilization month to several term thresholds and the duration threshold; The dimension evaluation data corresponding to the customer includes consumption limit dimension data, consumption frequency dimension data and usage duration dimension data, the consumption limit dimension data and consumption frequency dimension data are both in high or low states, and the usage duration dimension data is in active or dormant states; Compare the average expenditure amount data with the average expenditure amount threshold, and compare the installment amount to cash withdrawal amount data with the installment amount to cash withdrawal amount threshold; If the average expenditure amount data is less than or equal to the average expenditure amount threshold and the installment amount to cash withdrawal amount data is less than or equal to the installment amount to cash withdrawal amount threshold, the consumption limit dimension data is determined to be low, otherwise the consumption limit dimension data is determined to be high; Compare the monthly average mobilization times data with the monthly average mobilization times threshold, and compare the mobilization months to a number of periods data with the mobilization months to a number of periods threshold; If the monthly average usage frequency data is less than or equal to the monthly average usage frequency threshold comparison and the usage month to a certain period data is less than the usage month to a certain period threshold, then the consumption frequency dimension data is determined to be low, otherwise the consumption frequency dimension data is determined to be high; The duration data is compared with the duration threshold. If the duration data is greater than the duration threshold, the mobilization duration dimension data is determined to be dormant, otherwise it is active.

5. The credit card customer value stratification method based on the RFM model according to claim 4 is characterized in that: The priority of the average expenditure amount data comparison is higher than the priority of the installment amount to cash withdrawal amount data comparison; The priority of the comparison of the monthly average number of utilization times data is higher than the priority of the comparison of the utilization months and several term data.

6. The credit card customer value stratification method based on the RFM model according to claim 5 is characterized in that: The determining of value stratification data according to the dimension evaluation data includes: The value stratification data includes important value customers, important development customers, important retention customers, important attention customers and ordinary customers; If the usage duration dimension data is active and the consumption amount dimension data and consumption frequency dimension data are both high, then the customer is determined to be an important value customer; If the usage duration dimension data is active, and one of the consumption amount dimension data and the consumption frequency dimension data is high, then the customer is determined to be an important development customer; If the usage duration dimension data is dormant, and at least one of the consumption amount dimension data and the consumption frequency dimension data is high, the customer is determined to be an important customer to retain; If the consumption amount dimension data and consumption frequency dimension data are both low, and the usage duration dimension data is active, then the customer is determined to be an important customer; If the consumption amount dimension data and the consumption frequency dimension data are both low, and the usage duration dimension data is dormant, then the customer is determined to be an ordinary customer.

7. The credit card customer value stratification method based on the RFM model according to claim 6 is characterized in that: The important development customers include the first important development customer and the second important development customer; When the consumption amount dimension data is high, the customer is determined to be the first important development customer; When the consumption frequency dimension data is high, it is determined to be the second most important development customer.

8. The credit card customer value stratification method based on the RFM model according to claim 7 is characterized in that: The important retained customers include the first important retained customers, the second important retained customers and the third important retained customers; When both the consumption amount dimension data and the consumption frequency dimension data are high, the customer is determined to be the first important customer to be retained; When the consumption amount dimension data is high and the consumption frequency dimension data is low, the customer is determined to be the second most important customer to be retained; When the consumption amount dimension data is low and the consumption frequency dimension data is high, the customer is determined to be the third most important customer to be retained.

9. The credit card customer value stratification method based on the RFM model according to claim 8 is characterized in that: The step of obtaining historical risk data of a customer's credit card within a certain period, excluding data in the value stratification data whose historical risk data exceeds a preset risk threshold, and obtaining risk analysis data, and processing and outputting the risk analysis data includes: The historical risk data includes historical overdue days data on credit cards, and the risk threshold includes an overdue days threshold; Excluding data whose historical overdue days data is greater than or equal to the overdue days threshold in the value stratification data, and obtaining risk analysis data; Processing the risk analysis data to obtain excluded value stratification data, corresponding customer ratio data in the corresponding value stratification data, and corresponding income ratio data in the corresponding value stratification data; The excluded value stratification data, customer ratio data and income ratio data are output to identify credit card value customers.

10. A credit card customer value stratification device based on the RFM model, characterized in that: According to the method described in any one of claims 1 to 9, the device comprises: An acquisition module, the acquisition module is used to acquire credit card transaction information of a customer within a certain period; A multi-dimensional classification processing module, wherein the multi-dimensional classification processing performs multi-dimensional classification processing on the credit card transaction information to obtain index data; A comparison module, wherein the comparison module compares the indicator data with the indicator threshold value to determine the dimension evaluation data corresponding to the customer; A value stratification module, wherein the value stratification module determines value stratification data according to the dimension evaluation data; A risk output module obtains historical risk data of a customer's credit card within a certain period of time, excludes data in the value stratification data whose historical risk data exceeds a preset risk threshold, obtains risk analysis data, processes and outputs the risk analysis data to identify credit card value customers.

Images