A credit card differentiated traffic management method and system based on customer behavior

By analyzing customer consumption behavior data, calculating key indicators and dynamic adjustment factors, performing differentiated classifications, and formulating personalized traffic management strategies, solving the problem that traditional credit card management models cannot identify customer characteristics, and achieving flexible and accurate credit card management and Risk monitoring.

CN118710384BActive Publication Date: 2024-11-19BANK OF SHANGHAI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411179471.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-27
Publication Date
2024-11-19
Estimated Expiration
2044-08-27

AI Technical Summary

Technical Problem

The traditional credit card management model cannot accurately identify customers' consumption habits, preferences and abilities, resulting in insufficient flexibility and meticulous services, and the data processing process is fixed and cumbersome, affecting the bank's timely response to changes in customer data.

Method used

Differentiated traffic management method based on customer behavior, by obtaining consumption behavior data, analyzing customers' consumption habits, preferences and abilities, calculating key indicators, determining dynamic adjustment factors, performing differentiated classification, and formulating traffic management strategies based on classification results, in real time Monitor credit card usage.

Benefits of technology

It realizes personalized services to customers, improves the efficiency and accuracy of strategy formulation, can respond faster to changes in market and customer demand, timely discover and respond to potential risks, and ensure the safety of assets and funds.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118710384B_ABST
    Figure CN118710384B_ABST
Patent Text Reader

Abstract

The present invention provides a credit card differentiated flow management method and system based on customer behavior, which relates to the field of data processing technology. The method includes: identifying and analyzing the consumption habits, consumption preferences and consumption capabilities of customers according to consumption behavior data to obtain the analysis results of customer behavior; calculating the key indicators of each customer according to the historical consumption behavior data of the customer, including the average consumption amount and consumption frequency of each customer; for each customer, determining the dynamic adjustment factor according to the difference between the key indicators and the benchmark consumption behavior; classifying the customers according to the analysis results and the dynamic adjustment factors to obtain the classification results; and determining the corresponding flow management strategies for different categories of customers according to the classification results. The present invention can accurately identify the consumption habits, preferences and capabilities of customers.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a credit card differentiated flow management method and system based on customer behavior. Background Art

[0002] Under the traditional credit card management model, banks tend to adopt a relatively unified service strategy. This strategy means that although customers have different personal backgrounds, consumption habits and preferences, the content and form of the services provided by the bank are relatively fixed, which may not be flexible and detailed enough.

[0003] This traditional model also has some subtle shortcomings in data processing.

[0004] For example, the downloaded data may not fully reflect the individual characteristics of each customer, resulting in certain limitations in the bank's in-depth understanding of customer needs. At the same time, the data processing process may be relatively fixed and cumbersome, affecting the bank's timely response to changes in customer data. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a credit card differentiated traffic management method and system based on customer behavior, which can accurately identify customers' consumption habits, preferences and capabilities.

[0006] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0007] In a first aspect, a differentiated credit card traffic management method based on customer behavior is provided, the method comprising:

[0008] Obtain customers’ consumption behavior data, including consumption time, consumption location, consumption amount, and consumption type;

[0009] Based on consumer behavior data, identify and analyze customers’ consumption habits, preferences, and spending power to obtain analysis results of customer behavior;

[0010] Calculate key indicators for each customer based on their historical consumption behavior data, including average consumption amount and consumption frequency;

[0011] For each customer, a dynamic adjustment factor is determined based on the difference between key indicators and benchmark consumption behavior;

[0012] Based on the analysis results and dynamic adjustment factors, customers are classified into different categories to obtain classification results;

[0013] According to the classification results, determine the corresponding traffic management strategy for different categories of customers;

[0014] Customers' credit card usage is monitored and managed in real time according to traffic management strategies.

[0015] Furthermore, based on the consumption behavior data, identify and analyze the customer's consumption habits, consumption preferences and consumption capacity to obtain the analysis results of customer behavior, including:

[0016] Preprocessing the consumption behavior data to obtain preprocessed data;

[0017] Determine the number of clusters k, and determine k points from the preprocessed data as the initial cluster centers;

[0018] For each point in the preprocessed data, calculate its distance from each cluster center and assign it to the corresponding cluster center;

[0019] Recalculate the center point of each cluster and repeat the steps until the preset number of iterations is reached to obtain the clustering result;

[0020] According to the clustering results, customers are divided into k categories, each of which represents a group of customers with similar consumption behaviors.

[0021] Further, the consumption behavior data is preprocessed to obtain preprocessed data, including:

[0022] The consumption behavior data is processed through Filter to obtain filtered data, where Indicates at time Smoothed data value after filtering; Indicates at time The original consumption behavior data value; is the window size used to calculate the weighted sum of the current data; is the window size used to compute the weighted sum of past smoothed data; is a smoothing parameter, with a value between 0 and 1;

[0023] The filtered data is normalized to obtain preprocessed data.

[0024] Furthermore, for each point in the preprocessed data, the distance between it and each cluster center is calculated, including:

[0025] For each point in the preprocessed data, Calculate the distance between it and each cluster center; and is the summation index, which is used to traverse all feature dimensions. and From 1 to change; is the total number of features of the data point, indicating that the point and cluster centers The number of dimensions; and They are points and cluster centers timestamp; and is the attenuation factor; Indicates The weight of each feature; Indicates Features and Pearson correlation coefficient between the features; and Respectively indicate points and cluster centers In the The value of the feature dimension; Represents a data point in the preprocessed data With cluster center The distance between.

[0026] Furthermore, for each customer, a dynamic adjustment factor is determined based on the difference between the key indicators and the benchmark consumption behavior, including:

[0027] For each customer, based on the difference between key indicators and benchmark consumption behavior, Calculating Dynamic Adjustment Factors ;in, Indicates that the customer The actual value of each key indicator; Indicates Baseline values ​​of key indicators; Indicates The weight of each key indicator; Indicates the total number of key indicators; represents the adjustment factor; Indicates the average amount of spending by a customer over a period of time; Indicates the average amount of consumption by a customer in a period of time; Indicates the average consumption frequency of customers over a period of time; Indicates the average consumption frequency of a customer over a period of time.

[0028] Furthermore, based on the analysis results and dynamic adjustment factors, customers are differentiated and classified to obtain classification results, including:

[0029] For each customer, the consumption behavior data and the impact of the dynamic adjustment factor are represented as a multi-dimensional vector;

[0030] Calculate the weighted average of all customers in each dimension to obtain a baseline behavior vector representing the average consumption behavior under the current market environment and dynamic factors. The dimensions include consumption amount, consumption frequency, consumption type and dynamic adjustment factor.

[0031] For each customer's baseline behavior vector, calculate the weighted distance length between it and the baseline behavior vector according to the weight in the analysis result;

[0032] Sort all customers according to the weighted distance length to obtain the sorting result;

[0033] Analyze the customer distribution in the sorting results to determine the weighted distance threshold;

[0034] According to the weighted distance threshold, customers are divided into different groups to obtain classification results.

[0035] Furthermore, for each customer's baseline behavior vector, the weighted distance length between it and the baseline behavior vector is calculated according to the weight in the analysis result, including:

[0036] For each customer's baseline behavior vector, according to the weight in the analysis results, Calculate the weighted distance length between it and the benchmark behavior vector; Calculate the weighted distance length between it and the benchmark behavior vector; Wherein, Indicates The weighted distance length between the consumption behavior vector of a customer and the benchmark behavior vector; Indicates Customers in Consumption behavior data in 3 dimensions; Represents the baseline behavior vector in The values ​​in the dimensions; It is The original weights of the dimensions; and is the coefficient; The purchase frequency of the representative customer; is the index of the customer; is the index of the dimension; is the total number of dimensions.

[0037] Second, a differentiated credit card traffic management system based on customer behavior, including:

[0038] The acquisition module is used to obtain the customer's consumption behavior data, including consumption time, consumption location, consumption amount, and consumption type; based on the consumption behavior data, identify and analyze the customer's consumption habits, consumption preferences, and consumption capacity to obtain the analysis results of the customer's behavior;

[0039] The processing module is used to calculate the key indicators of each customer based on the customer's historical consumption behavior data, including the average consumption amount and consumption frequency of each customer; for each customer, determine the dynamic adjustment factor based on the difference between the key indicators and the benchmark consumption behavior; based on the analysis results and the dynamic adjustment factor, differentiate the customers to obtain the classification results; based on the classification results, determine the corresponding traffic management strategy for different categories of customers; based on the traffic management strategy, monitor and manage the customer's credit card use in real time.

[0040] According to a third aspect, a computing device includes:

[0041] one or more processors;

[0042] The storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the above method.

[0043] In a fourth aspect, a computer-readable storage medium stores a program, and the program implements the above method when executed by a processor.

[0044] The above solution of the present invention includes at least the following beneficial effects:

[0045] By obtaining the consumption behavior data of customers, including detailed information such as consumption time, location, amount and type, the comprehensive integration of data is achieved. Through the analysis of consumption behavior, this method can accurately identify the consumption habits, preferences and capabilities of customers. Based on the calculation of key indicators of each customer, such as average consumption amount and consumption frequency, the bank can tailor dynamic adjustment factors for each customer to achieve personalized services.

[0046] Through differentiated classification, banks can quickly divide customers into different groups and formulate corresponding flow management strategies for each group. This classification method greatly improves the efficiency and accuracy of strategy formulation, allowing banks to respond more quickly to market changes and changes in customer needs. According to the flow management strategy, banks can monitor and manage customers' credit card usage in real time, which not only helps to timely discover and respond to potential risks such as abnormal transactions or fraudulent behavior, but also ensures the safety of bank assets and customer funds. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flow chart of a differentiated credit card traffic management method based on customer behavior provided by an embodiment of the present invention.

[0048] Figure 2 It is a schematic diagram of a differentiated credit card traffic management system based on customer behavior provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0050] like Figure 1 As shown, an embodiment of the present invention proposes a credit card differentiated flow management method based on customer behavior, the method comprising:

[0051] Step 11, obtaining the customer's consumption behavior data, including consumption time, consumption location, consumption amount, and consumption type;

[0052] Step 12: Identify and analyze the customer's consumption habits, consumption preferences, and consumption capacity based on the consumption behavior data to obtain the analysis results of the customer's behavior;

[0053] Step 13, based on the customer's historical consumption behavior data, calculate each customer's key indicators, including each customer's average consumption amount and consumption frequency;

[0054] Step 14, for each customer, determine a dynamic adjustment factor based on the difference between the key indicator and the benchmark consumption behavior;

[0055] Step 15, based on the analysis results and the dynamic adjustment factors, differentiate the customers to obtain classification results;

[0056] Step 16, according to the classification results, determine the corresponding traffic management strategy for different categories of customers;

[0057] Step 17: Monitor and manage the customer's credit card usage in real time according to the traffic management strategy.

[0058] In an embodiment of the present invention, step 11, by collecting the consumption behavior data of the customer, provides a rich data source for subsequent data analysis and customer behavior profiling. Step 12, through in-depth analysis of consumption data, the customer's consumption behavior and preferences can be understood more accurately. Step 13, by calculating key indicators, it is helpful to predict the customer's future behavior, thereby optimizing resource allocation and service strategies. Step 14, the introduction of dynamic adjustment factors makes the traffic management strategy more flexible and personalized. By comparing customer behavior with benchmark behavior, the service strategy can be adjusted more accurately to meet the personalized needs of different customers. Step 15, through data-driven differentiated classification, the needs and characteristics of different customer groups can be more effectively identified. Step 16, data-based traffic management strategies can ensure efficient use of resources. Step 17, real-time monitoring can promptly detect and handle abnormal situations to ensure the safety of customer funds. At the same time, through continuous tracking and analysis of customer behavior, service quality and efficiency are improved.

[0059] Step 11 involves collecting various consumption data generated by customers when using credit cards. Record the specific time when customers use credit cards to make transactions, which can be year, month, day, hour, minute, or even second. Consumption time data helps analyze customers' consumption habits and active periods. For example, some customers may prefer to consume on weekends or at night. Record the location information of customers' credit card transactions, which can be specific to the merchant name, address, or location data obtained through GPS positioning. By analyzing the consumption location, you can understand the customer's activity range and preferences, such as whether they prefer to consume in a specific area or type of merchant. Record the specific amount of each transaction. The consumption amount is an important indicator for evaluating customers' consumption capacity and consumption habits. For example, large transactions may indicate a customer's preference for high-end products or services. Indicate the category to which the transaction belongs, such as catering, shopping, entertainment, travel, etc. Consumption type data helps identify customers' interests and consumption preferences.

[0060] For example, suppose a bank implements a credit card differentiated flow management project based on customer behavior. During the project launch phase, the bank collects a large amount of customer consumption behavior data through its credit card system.

[0061] Take one of our customers as an example, some of his consumption data is as follows:

[0062] Consumption time: customers often consume on weekend nights, especially on Friday and Saturday nights;

[0063] In terms of consumption locations, data showed that customers frequently visited a fine-dining restaurant in the city center and a nearby movie theater.

[0064] Amount of consumption: the average amount spent in this restaurant is 500 yuan, while the average amount spent in the cinema is 100 yuan.

[0065] In terms of consumption type, customers’ consumption is mainly concentrated in two categories: catering and entertainment.

[0066] Based on these data, the bank can preliminarily infer that the customer may be a customer who values ​​quality of life and likes to engage in leisure and entertainment activities on weekends. Therefore, in the subsequent traffic management strategy, the bank can consider providing customers with more preferential activities or value-added services related to catering and entertainment to enhance customer stickiness and satisfaction. At the same time, by analyzing more data of similar customers, the bank can also further optimize its credit card products and services to meet the personalized needs of different customer groups.

[0067] In a preferred embodiment of the present invention, the above step 12, identifying and analyzing the customer's consumption habits, consumption preferences and consumption capacity according to the consumption behavior data to obtain the analysis results of the customer behavior, may include:

[0068] Step 121, preprocessing the consumption behavior data to obtain preprocessed data;

[0069] Step 122, determining the number of clusters k, and determining k points from the preprocessed data as initial cluster centers;

[0070] Step 123, for each point in the preprocessed data, calculate its distance from each cluster center and assign it to the corresponding cluster center;

[0071] Step 124, recalculate the center point of each cluster, and repeat the steps until a preset number of iterations is reached to obtain a clustering result;

[0072] Step 125 , according to the clustering result, the customers are divided into k categories, each category represents a group of customers with similar consumption behaviors.

[0073] In an embodiment of the present invention, step 121, data preprocessing can clean and standardize the original data, eliminate outliers, missing values ​​and duplicate values, so that subsequent data analysis is more accurate and effective. Preprocessing can also convert data into a format suitable for cluster analysis, thereby improving the accuracy and efficiency of clustering. Step 122, determine the number of clusters k, and by selecting a suitable k value, ensure that the clustering results are neither too rough nor too fragmented, thereby better reflecting the actual consumer behavior groups of customers. At the same time, selecting a suitable initial cluster center can accelerate the convergence of the clustering process. Step 123, by calculating the distance between each data point and each cluster center and assigning it to the nearest cluster center, it can be ensured that the data points within each cluster have the greatest similarity. Step 124, by continuously iteratively calculating the center point of each cluster, the clustering results can be made more stable and accurate. As the number of iterations increases, the cluster center will gradually approach the real customer consumer behavior center, thereby improving the quality of clustering. Step 125, divide the customers into different categories according to the clustering results, and each category represents a group of customers with similar consumer behavior.

[0074] When applied specifically, step 121, data preprocessing is an important step before data analysis, and its purpose is to ensure the quality and consistency of the data to improve the accuracy of subsequent analysis. In this step, a series of processing operations will be performed on the original consumer behavior data. After preprocessing, the obtained data set will be more standardized and neat, providing a good foundation for subsequent clustering analysis.

[0075] Step 122, in cluster analysis, it is crucial to determine the appropriate number of clusters k. The choice of k value directly affects the precision and interpretability of the clustering results. In order to determine the k value, the silhouette coefficient method can be used. After determining the k value, k points are randomly or according to certain strategies selected from the preprocessed data as the initial cluster centers. These center points are the starting points for subsequent cluster iterations. Step 123, in this step, each data point in the preprocessed data will be traversed, and the Euclidean distance between the point and each cluster center will be calculated. According to the calculated distance, the data points are assigned to the category where the nearest cluster center is located. In this way, each data point will be classified into a specific cluster. Step 124, after the data points are initially assigned to the cluster centers, the center points of each cluster need to be recalculated. The new center point is usually the average or median of all data points in the cluster. Then, based on the new center point, step 123 is performed again to redistribute the data points to the nearest cluster center. This process will be iterated continuously until the preset number of iterations is reached. After multiple iterations, the cluster centers will gradually stabilize and form the final clustering results. Step 125, after the above cluster analysis process, k stable cluster centers and their corresponding data point sets will be obtained. Each cluster represents a group of customers with similar consumption behaviors. In this step, customers will be clearly divided into k categories based on the clustering results.

[0076] In a preferred embodiment of the present invention, the above step 121, preprocessing the consumption behavior data to obtain preprocessed data, may include:

[0077] Step 1211: the consumption behavior data is processed by Filter to obtain filtered data, where Indicates at time Smoothed data value after filtering; Indicates at time The original consumption behavior data value; is the window size used to calculate the weighted sum of the current data; is the window size used to compute the weighted sum of past smoothed data; is a smoothing parameter, with a value between 0 and 1;

[0078] The filtered data is normalized to obtain preprocessed data.

[0079] Step 1212, normalize the filtered data to obtain preprocessed data.

[0080] In the embodiment of the present invention, step 1211, by applying a given filtering formula, can effectively smooth the short-term fluctuations and outliers in the original consumer behavior data, thereby more clearly revealing the long-term trend and periodic changes of the data. Filtering can reduce the impact of random noise on the data and improve the quality and reliability of the data; by adjusting the smoothing parameter , Window Size and , the degree of data smoothing can be flexibly controlled to meet different analysis requirements. Step 1212, standardization processing can eliminate the impact of different dimensions of the original data on the analysis results, so that data of different dimensions can be compared and analyzed on the same scale. Standardized data can reduce the model's dependence on specific dimensions, thereby improving the model's stability and generalization ability.

[0081] In a preferred embodiment of the present invention, the above step 123, for each point in the preprocessed data, calculating the distance between it and each cluster center, may include:

[0082] Step 1231, for each point in the preprocessed data, Calculate the distance between it and each cluster center; and is the summation index, which is used to traverse all feature dimensions. and From 1 to change; is the total number of features of the data point, indicating that the point and cluster centers The number of dimensions; and They are points and cluster centers timestamp; and is the attenuation factor; Indicates The weight of the feature; Indicates Features and Pearson correlation coefficient between the features; and Respectively indicate points and cluster centers In the The value of the feature dimension; Represents a data point in the preprocessed data With cluster center The distance between.

[0083] In the embodiment of the present invention, by introducing feature weights , this formula can adjust the contribution of different features to the distance calculation according to their importance. This helps to more accurately reflect the influence of different features in the clustering process. By introducing the time factor and , and the attenuation factor , this formula can take into account the time attribute of data points, making it easier for points that are closer in time to be classified into one category when clustering. By introducing the Pearson correlation coefficient , which takes into account the correlation between features. This helps to more accurately measure the similarity between data points, because highly correlated features will be appropriately adjusted in the distance calculation. This formula provides a high degree of flexibility and can adapt to different data characteristics and analysis needs by adjusting parameters such as feature weights, decay factors, etc.

[0084] Step 13 specifically includes:

[0085] Based on each customer's historical consumption behavior data, calculate the corresponding average consumption amount and consumption frequency. Calculate the average consumption amount of each customer:

[0086] First, it is necessary to collect all consumption records of customers in the past period of time.

[0087] Next, the consumption amount of each customer is accumulated.

[0088] Then, count the number of times each customer consumes.

[0089] Finally, divide each customer's total spending amount by the number of times they spend to get the average spending amount.

[0090] Calculate the consumption frequency of each customer:

[0091] Consumption frequency refers to the number of times a customer consumes within a certain period of time. The consumption frequency can be obtained by counting the number of times each customer consumes within a specified period of time.

[0092] The calculation results of these two indicators can facilitate understanding of customers' consumption behavior and consumption capacity.

[0093] For example, suppose you are analyzing the consumption behavior of credit card customer B in order to provide more accurate services and recommendations. Now, you will calculate the average consumption amount and consumption frequency of customer B based on his historical credit card consumption data.

[0094] Extract customer B's credit card spending records for the past year (assuming 12 months) from the bank system. Filter and sort out the monthly spending amount and number of spending.

[0095] Case Study:

[0096] Calculation of average consumption amount:

[0097] Assume that customer B's total consumption in the past 12 months is RMB 12,000; during these 12 months, customer B made a total of 30 transactions, so the average transaction amount is: RMB 12,000 ÷ 30 times = RMB 400 per transaction.

[0098] Calculation of consumption frequency:

[0099] In the past 12 months, Customer B has made 30 transactions; therefore, the average number of transactions per month is: 30 ÷ 12 months = 2.5 / month.

[0100] The average amount of money spent by customer B each time using his credit card is RMB 400, indicating that he has a certain spending power. He uses his credit card an average of 2.5 times per month, indicating that he is a relatively active credit card user. Based on the above analysis, the bank can consider recommending higher-end credit card services to customer B, or providing promotional activities that match his spending habits. At the same time, the bank can also strengthen communication with customer B to understand more of his needs in order to provide more personalized financial services.

[0101] In a preferred embodiment of the present invention, the above step 14, for each customer, determining the dynamic adjustment factor according to the difference between the key indicator and the benchmark consumption behavior, may include:

[0102] For each customer, based on the difference between key indicators and benchmark consumption behavior, Calculating Dynamic Adjustment Factors ;in, Indicates that the customer The actual value of each key indicator; Indicates Baseline values ​​of key indicators; Indicates The weight of each key indicator; Indicates the total number of key indicators; represents the adjustment factor; Indicates the average amount of spending by a customer over a period of time; Indicates the average amount of consumption by a customer in a period of time; Indicates the average consumption frequency of customers over a period of time; Indicates the average consumption frequency of a customer over a period of time.

[0103] In an embodiment of the present invention, by calculating the dynamic adjustment factor of each customer, the enterprise can formulate personalized service strategies for the consumption behaviors of different customers. This personalized service helps to improve customer satisfaction because the service is more in line with the actual needs and consumption habits of customers. The calculation method of the dynamic adjustment factor takes into account multiple key indicators, including the difference between the actual consumption value of the customer and the benchmark value, the weight of each key indicator, and the adjustment coefficient. This comprehensive consideration makes the dynamic adjustment factor highly flexible and can adapt to the diverse consumption needs of different customers. By comparing the difference between the average consumption amount and consumption frequency of the customer and the benchmark value, the consumption preference and consumption capacity of the customer can be more accurately identified. The application of the dynamic adjustment factor is not limited to the adjustment of service strategies, but can also be used to optimize customer relationship management. By regularly calculating and updating the dynamic adjustment factor of the customer, changes in customer needs can be discovered in a timely manner, thereby adjusting the interaction method and frequency with the customer and maintaining close contact with the customer.

[0104] In a preferred embodiment of the present invention, the above step 15, based on the analysis results and the dynamic adjustment factors, classifying the customers into different categories to obtain classification results, may include:

[0105] Step 151, for each customer, the consumption behavior data and the impact of the dynamic adjustment factor are expressed as a multi-dimensional vector;

[0106] Step 152, calculating the weighted average of all customers in each dimension to obtain a baseline behavior vector representing the average consumption behavior under the current market environment and the influence of dynamic factors, wherein each dimension includes consumption amount, consumption frequency, consumption type and dynamic adjustment factor;

[0107] Step 153, for each customer's baseline behavior vector, calculate the weighted distance length between it and the baseline behavior vector according to the weight in the analysis result;

[0108] Step 154, sorting all customers according to the weighted distance length to obtain a sorting result;

[0109] Step 155, analyzing the distribution of customers in the sorting results to determine a weighted distance threshold;

[0110] Step 156, classify the customers into different groups according to the weighted distance threshold to obtain a classification result.

[0111] In an embodiment of the present invention, through comprehensive analysis of multi-dimensional consumer behavior and dynamic adjustment factors, enterprises can more accurately identify the characteristics and behavior patterns of different customer groups. Based on the differentiated classification of customers, more personalized marketing strategies can be formulated. For different types of customers, product or service information that is more in line with their interests and needs can be pushed to improve marketing effectiveness and customer satisfaction. By classifying customers, marketing resources can be allocated more reasonably. In the field of financial services, especially in bank credit card business, differentiated classification of customers also helps in risk management and prediction. By analyzing the consumer behavior and credit records of different categories of customers, potential risk points can be discovered in a timely manner, and corresponding prevention and control measures can be taken.

[0112] In a preferred embodiment of the present invention, the above step 153, for each customer's baseline behavior vector, calculates the weighted distance length between the baseline behavior vector and the baseline behavior vector according to the weight in the analysis result, including:

[0113] Step 1531: for each customer's baseline behavior vector, according to the weight in the analysis result, Calculate the weighted distance length between it and the benchmark behavior vector; where, Indicates The weighted distance length between the consumption behavior vector of a customer and the benchmark behavior vector; Indicates Customers in Consumption behavior data in 3 dimensions; Represents the baseline behavior vector in The values ​​in the dimensions; It is The original weights of the dimensions; and is the coefficient; The purchase frequency of the representative customer; is the index of the customer; is the index of the dimension; is the total number of dimensions.

[0114] In the embodiment of the present invention, by comprehensively considering the consumption behavior data of multiple dimensions (such as consumption amount, consumption frequency, consumption type, etc.) and the influence of dynamic adjustment factors, the calculation of weighted distance length can more comprehensively and accurately evaluate the consumption behavior of customers. The calculation method of weighted distance length takes into account the weights of different dimensions, which means that different consumption behavior characteristics have different importance in the evaluation. This personalized measurement method can more accurately reflect the consumption characteristics and preferences of customers, thereby providing enterprises with more accurate customer portraits. By introducing the coefficient and , and the customer's purchase frequency , the calculation of weighted distance length has higher flexibility. As an important basis for customer classification, the accuracy of weighted distance length directly affects the results of customer classification. By comprehensively considering multiple factors and dimensions, the calculation of weighted distance length can improve the accuracy of customer classification and help enterprises identify different types of customers more accurately.

[0115] When specifically applied, this includes:

[0116] Step 151, for each customer in the database, extract their consumption behavior data, including but not limited to: consumption amount, consumption frequency, consumption type; combine each customer's consumption behavior data and its corresponding dynamic adjustment factor into a multi-dimensional vector. For example, a customer's vector may be in the form of: [consumption amount, consumption frequency, consumption type 1, consumption type 2, ..., dynamic adjustment factor].

[0117] Step 152, traverse the consumption behavior vectors of all customers; for each dimension (such as consumption amount, consumption frequency, etc.), calculate the weighted average of all customers in this dimension, where the weight can be set according to historical data. The weighted averages of each dimension are combined into a new vector, which is the baseline behavior vector, which represents the average consumption behavior under the current market environment and dynamic factors.

[0118] Step 153, initialize an empty list to store the weighted distance length; for each customer's consumption behavior vector, perform the following operations:

[0119] Initialize the accumulated value of the weighted distance length to 0; traverse each dimension of the customer's consumption behavior vector; for each dimension, calculate the difference between the value on that dimension and the value of the corresponding dimension of the benchmark behavior vector, and square it; multiply the squared difference by the weight of the dimension and the adjustment factor (purchase frequency and dynamic adjustment factor ); add the weighted square difference to the weighted distance length; take the square root of the accumulated value to obtain the weighted distance length between the customer and the benchmark behavior vector, and add it to the previously initialized list.

[0120] Step 154, sorting the weighted distance length list calculated in step 153, and recording the corresponding customer index.

[0121] Step 155, analyzing the sorted weighted distance length list, one or more thresholds may be determined by observing the distribution of data, setting fixed thresholds, or using a clustering algorithm.

[0122] Step 156, according to the weighted distance threshold determined in step 155, the customers are divided into different groups. For example, customers whose weighted distance is less than a certain threshold can be set as "core customers", customers whose weighted distance is greater than another threshold can be set as "marginal customers", and customers between the two can be set as "potential customers". The classification results can be output, and different marketing strategies and service plans can be formulated for each type of customer.

[0123] Step 16 specifically includes:

[0124] After completing the differentiated classification of customers, formulate appropriate traffic management strategies for different types of customers. Traffic management strategies refer to the management and restriction measures for the frequency, amount, type, etc. of customer credit card transactions, aiming to ensure transaction security, control risks, and improve customer satisfaction.

[0125] Specifically, for "core customers", banks may provide more relaxed traffic management strategies, such as increasing the single transaction limit, relaxing the daily cumulative transaction limit, etc., because these customers usually have good credit records and are important customers of the bank. For "marginal customers" or customers with higher risks, banks may adopt stricter traffic management strategies, such as reducing transaction limits, increasing transaction verification steps, etc., to control potential risks.

[0126] Step 17 specifically includes:

[0127] After formulating traffic management strategies for different customers, banks need to establish a real-time monitoring system to track customers' credit card usage. This system can monitor transaction activities in real time. Once abnormal transactions are detected (such as exceeding the limit, transactions at abnormal locations or times, etc.), the system will trigger an alarm and take corresponding management measures, such as suspending transactions, requiring customers to confirm transactions, etc.

[0128] Specific examples:

[0129] Suppose that a bank classifies its customers into three categories: "core customers", "ordinary customers" and "marginal customers" through customer differentiation.

[0130] For Mr. A, a "core customer", the bank has formulated a relatively loose flow management strategy for him because he is a high-end customer of the bank with an excellent credit record, such as increasing the single transaction limit to 50,000 yuan, the daily cumulative transaction limit to 200,000 yuan, and providing 24-hour customer service support. In daily use, Mr. A's transaction behavior is less restricted and he can use his credit card more freely.

[0131] For Ms. B, an "ordinary customer", the bank adopted a standard flow management strategy, such as a single transaction limit of 20,000 yuan and a daily cumulative transaction limit of 100,000 yuan. Her transaction behavior is monitored to a certain extent, but it will not trigger an alarm when used within the normal range.

[0132] As for Mr. C, a "marginal customer", the bank has adopted strict traffic management strategies due to his poor credit record or other risk factors, such as a single transaction limit of 5,000 yuan and each transaction needs to be verified by SMS. At the same time, the bank has carried out more stringent real-time monitoring of his transaction behavior. Once Mr. C's transaction behavior is abnormal (such as multiple consecutive transactions in a short period of time, sudden changes in transaction locations, etc.), the bank's monitoring system will immediately trigger an alarm and may suspend his credit card transactions until the transaction security is confirmed with the customer.

[0133] Through such classification management and real-time monitoring, banks can provide more personalized services and improve customer satisfaction while ensuring transaction security.

[0134] like Figure 2 As shown, the embodiment of the present invention also provides a credit card differentiated flow management system 20 based on customer behavior, including:

[0135] The acquisition module 21 is used to acquire the consumption behavior data of the customer, including consumption time, consumption location, consumption amount, and consumption type; based on the consumption behavior data, identify and analyze the consumption habits, consumption preferences, and consumption capacity of the customer to obtain the analysis results of the customer behavior;

[0136] Processing module 22 is used to calculate the key indicators of each customer based on the customer's historical consumption behavior data, including the average consumption amount and consumption frequency of each customer; for each customer, determine the dynamic adjustment factor based on the difference between the key indicators and the benchmark consumption behavior; based on the analysis results and the dynamic adjustment factor, differentiate the customers to obtain classification results; based on the classification results, determine the corresponding traffic management strategy for different categories of customers; based on the traffic management strategy, monitor and manage the customer's credit card use in real time.

[0137] It should be noted that the system is a system corresponding to the above method, and all implementation methods in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0138] The embodiment of the present invention further provides a computing device, comprising: a processor, a memory storing a computer program, wherein when the computer program is executed by the processor, the method described above is executed. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0139] The embodiment of the present invention also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the method described above. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0140] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A credit card differentiated traffic management method based on customer behavior, characterized in that: The method comprises: Obtain customers’ consumption behavior data, including consumption time, consumption location, consumption amount, and consumption type; Based on consumer behavior data, identify and analyze customers’ consumption habits, preferences, and spending power to obtain analysis results of customer behavior; Calculate key indicators for each customer based on their historical consumption behavior data, including average consumption amount and consumption frequency; For each customer, a dynamic adjustment factor is determined based on the difference between key indicators and benchmark consumption behavior; Based on the analysis results and dynamic adjustment factors, customers are classified into different categories to obtain classification results; According to the classification results, determine the corresponding traffic management strategy for different categories of customers; According to the traffic management strategy, the customer's credit card usage is monitored and managed in real time; based on the consumption behavior data, the customer's consumption habits, consumption preferences and consumption capacity are identified and analyzed to obtain the analysis results of customer behavior, including: Preprocessing the consumption behavior data to obtain preprocessed data; Determine the number of clusters k, and determine k points from the preprocessed data as the initial cluster centers; For each point in the preprocessed data, calculate its distance from each cluster center and assign it to the corresponding cluster center; Recalculate the center point of each cluster and repeat the steps until the preset number of iterations is reached to obtain the clustering result; According to the clustering results, the customers are divided into k categories, each category represents a group of customers with similar consumption behaviors; the consumption behavior data is preprocessed to obtain preprocessed data, including: The consumption behavior data is processed through Filter to obtain filtered data, where y t represents the smoothed data value after filtering at time t; x i represents the original consumption behavior data value at time i; n is the window size used to calculate the weighted sum of current data; m is the window size used to calculate the weighted sum of past smoothed data; α is the smoothing parameter, which takes a value between 0 and 1; The filtered data is standardized to obtain preprocessed data; for each point in the preprocessed data, the distance between it and each cluster center is calculated, including: For each point in the preprocessed data, Calculate its distance from each cluster center; Among them, b and g are summation indicators used to traverse all feature dimensions, b and g vary from 1 to n1; n1 is the total number of features of the data point, indicating the number of dimensions of point p and cluster center c; t p and t c are the timestamps of point p and cluster center c respectively; α1 and β are decay factors; w b represents the weight of the bth feature; R bg represents the Pearson correlation coefficient between the bth feature and the gth feature; p b and c b Represent the values ​​of point p and cluster center p in the bth feature dimension respectively; d(p,c) represents the distance between a data point p and cluster center c in the preprocessed data; For each customer, the dynamic adjustment factor is determined according to the difference between the key indicators and the benchmark consumption behavior, including: For each customer, based on the difference between key indicators and benchmark consumption behavior, Calculate the dynamic adjustment factor D AF ; Among them, K s Indicates the actual value of the customer on the sth key indicator; B s represents the benchmark value of the sth key indicator; W s represents the weight of the sth key indicator; u represents the total number of key indicators; α2 represents the adjustment coefficient; A r Indicates the average amount of consumption by a customer over a period of time; A h Indicates the average amount of consumption by customers in a period of time; F r Indicates the average consumption frequency of customers in a period of time; F h Indicates the average consumption frequency of customers over a period of time; based on the analysis results and dynamic adjustment factors, customers are differentiated and classified to obtain classification results, including: For each customer, the consumption behavior data and the impact of the dynamic adjustment factor are represented as a multi-dimensional vector; Calculate the weighted average of all customers in each dimension to obtain a baseline behavior vector representing the average consumption behavior under the current market environment and dynamic factors. The dimensions include consumption amount, consumption frequency, consumption type and dynamic adjustment factor. For each customer's baseline behavior vector, calculate the weighted distance length between it and the baseline behavior vector according to the weight in the analysis result; Sort all customers according to the weighted distance length to obtain the sorting result; Analyze the customer distribution in the sorting results to determine the weighted distance threshold; According to the weighted distance threshold, customers are divided into different groups to obtain classification results; for each customer's baseline behavior vector, the weighted distance length between it and the baseline behavior vector is calculated according to the weight in the analysis result, including: For each customer's baseline behavior vector, according to the weight in the analysis results, Calculate the weighted distance length between it and the benchmark behavior vector; in, Indicates The weighted distance length between the consumption behavior vector of a customer and the benchmark behavior vector; Indicates Customers in Consumption behavior data in 3 dimensions; Denotes the baseline behavior vector in The value of the dimension; It is The original weights of the dimensions; α3 and β2 are coefficients; C p The purchase frequency of the representative customer; is the index of the customer; is the index of the dimension; μ is the total number of dimensions.

2. A credit card differentiated flow management system based on customer behavior, characterized in that: Applied in the method as claimed in claim 1, comprising: The acquisition module is used to obtain the customer's consumption behavior data, including consumption time, consumption location, consumption amount, and consumption type; based on the consumption behavior data, identify and analyze the customer's consumption habits, consumption preferences, and consumption capacity to obtain the analysis results of the customer's behavior; The processing module is used to calculate the key indicators of each customer based on the customer's historical consumption behavior data, including the average consumption amount and consumption frequency of each customer; for each customer, determine the dynamic adjustment factor based on the difference between the key indicators and the benchmark consumption behavior; based on the analysis results and the dynamic adjustment factor, differentiate the customers to obtain the classification results; based on the classification results, determine the corresponding traffic management strategy for different categories of customers; based on the traffic management strategy, monitor and manage the customer's credit card use in real time.

3. A computing device, characterized in that: include: one or more processors; A storage device, used for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method as claimed in claim 1.

4. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, which implements the method according to claim 1 when executed by a processor.