Bank customer classification method and device, electronic equipment and computer readable storage medium

Abstract

The invention discloses a bank customer classification method and device, electronic equipment and a computer readable storage medium. The method comprises the following steps: acquiring customer data of customers to be classified; obtaining first feature data related to the asset level change, and obtaining an asset level improvement probability prediction value according to the first feature data; performing primary prediction on the asset marketing potential of the to-be-classified customers according to the asset level promotion probability prediction value; when the to-be-classified customers are predicted to have the asset marketing potential for the first time, determining that the to-be-classified customers belong to a first type of asset marketing customer group; when it is predicted that the to-be-classified customers do not have the asset marketing potential for the first time, obtaining second feature data related to the asset potential according to the customer data, obtaining an asset potential predicted value of the to-be-classified customers according to the second feature data, and predicting the asset marketing potential of the to-be-classified customers again according to the asset potential predicted value, and determining an asset marketing customer group to which the to-be-classified customers belong according to a re-prediction result.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a bank customer classification method, device, electronic device and computer-readable storage medium. Background Art

[0002] In the banking business scenario, the bank has a huge number of customers. If a suitable asset marketing plan is formulated for each customer separately, a large amount of system resources will be consumed, resulting in low system efficiency. Based on this, the existing technology uses bank customer data as a basis to classify similar customers into a customer group. The bank system determines the asset marketing strategy suitable for the customer group based on the characteristics of the customer group, and recommends assets to all customers in the customer group based on the asset marketing strategy. Compared with formulating a suitable asset marketing plan for each user separately, it can reduce the system's work content, thereby reducing the system's resource consumption.

[0003] However, the current customer classification scheme is rather crude, resulting in poor asset marketing results. Summary of the invention

[0004] In view of the defects in customer classification in the prior art and to at least solve some of the technical problems therein, the embodiments of the present application provide a bank customer classification method, device, electronic device and computer-readable storage medium.

[0005] The present application embodiment adopts the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a bank customer classification method, the bank customer classification method comprising:

[0007] Obtain customer data of the customer to be classified;

[0008] Acquire first characteristic data related to asset grade change according to the customer data, and acquire a predicted value of the probability of asset grade improvement of the customer to be classified according to the first characteristic data;

[0009] Performing an initial forecast of the asset marketing potential of the customer to be classified according to the asset grade improvement probability forecast value;

[0010] When initially predicting that the customer to be classified has asset marketing potential, determining that the customer to be classified belongs to the first type of asset marketing customer group;

[0011] When it is initially predicted that the customer to be classified does not have asset marketing potential, second characteristic data related to asset potential is obtained based on the customer data, and a predicted value of the asset potential of the customer to be classified is obtained based on the second characteristic data, and the asset marketing potential of the customer to be classified is re-predicted based on the predicted value of asset potential, and the asset marketing customer group to which the customer to be classified belongs is determined based on the re-prediction result.

[0012] In a second aspect, an embodiment of the present application further provides a bank customer classification device, the bank customer classification device comprising:

[0013] A data acquisition unit, used to acquire customer data of customers to be classified;

[0014] A data processing unit, configured to obtain first characteristic data related to asset grade change according to the customer data, and obtain a predicted value of the probability of asset grade improvement of the customer to be classified according to the first characteristic data;

[0015] A prediction and classification unit is used to make an initial prediction of the asset marketing potential of the customer to be classified according to the predicted value of the probability of asset grade improvement, and when it is initially predicted that the customer to be classified has the asset marketing potential, determine that the customer to be classified belongs to the first type of asset marketing customer group; and when it is initially predicted that the customer to be classified does not have the asset marketing potential, obtain second characteristic data related to the asset potential according to the customer data, and obtain the predicted value of the asset potential of the customer to be classified according to the second characteristic data, re-predict the asset marketing potential of the customer to be classified according to the predicted value of the asset potential, and determine the asset marketing customer group to which the customer to be classified belongs according to the re-prediction result.

[0016] In a third aspect, an embodiment of the present application further provides an electronic device, including:

[0017] a processor; and a memory arranged to store computer executable instructions which, when executed, cause the processor to perform a bank customer classification method.

[0018] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs, and when the one or more programs are executed by an electronic device including multiple application programs, the electronic device executes a bank customer classification method.

[0019] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects:

[0020] The embodiment of the present application first uses the first characteristic data to obtain the predicted value of the probability of asset grade improvement of the customer to be classified, and makes an initial prediction of the asset marketing potential of the customer to be classified based on the predicted value of the probability of asset grade improvement. When it is initially predicted that the customer to be classified has the asset marketing potential, it is determined that the customer to be classified belongs to the first type of asset marketing customer group; when it is initially predicted that the customer to be classified does not have the asset marketing potential, the first characteristic data is used to obtain the predicted value of the asset potential of the customer to be classified, and the asset marketing potential of the customer to be classified is predicted again based on the predicted value of the asset potential, so as to determine the asset marketing customer group of the customer to be classified in combination with the predicted value of the probability of asset grade improvement and the predicted value of the asset potential, thereby improving the accuracy of customer classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0022] Figure 1 A schematic diagram of a flow chart of a bank customer classification method in an embodiment of the present application;

[0023] Figure 2 A schematic diagram of a process for classifying the asset marketing potential of bank customers in an embodiment of the present application;

[0024] Figure 3 This is a schematic diagram of an asset marketing process based on the classification of bank customers in an embodiment of the present application;

[0025] Figure 4 This is a schematic diagram of the structure of a bank customer classification device in an embodiment of the present application;

[0026] Figure 5 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0028] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0029] The present application embodiment provides a bank customer classification method, such as Figure 1As shown, a flow chart of a bank customer classification method in an embodiment of the present application is provided, and the method at least includes the following steps S110 to S150:

[0030] Step S110, obtaining customer data of the customer to be classified.

[0031] The executor of the bank customer classification method of the embodiment of the present application is a computing device. When the computing device classifies the customers to be classified, it first obtains the customer data of the customers to be classified. The customer data includes basic customer information (such as gender, age, education level, occupation code, residential address, travel information, etc.), asset data (such as credit card consumption amount, average daily assets in recent months, total assets, total loans, contracted and held product information, capital flow, etc.), transaction data (such as account transaction information, medium- and long-term transaction information, RFM (Recency-Frequency-Monetary) information, etc.), etc.

[0032] It is understandable that the above computing device can be hardware or software. When the computing device is hardware, it can be a distributed cluster consisting of multiple servers or terminal devices, or a single server or a single terminal device. When the computing device is embodied as software, it can be installed in the hardware devices listed above. The computing device can also be multiple software or software modules used to provide distributed services, or it can be implemented as a single software or software module, which is not specifically limited here.

[0033] Step S120: acquiring first characteristic data related to asset grade change according to the customer data, and acquiring a predicted value of the probability of asset grade improvement of the customer to be classified according to the first characteristic data.

[0034] This embodiment processes customer data through a certain data processing strategy to obtain first feature data. The first feature data at this time may include feature data that is positively correlated with asset level improvement, and may also include feature data that is negatively correlated with asset level improvement. The first feature data is inferred and predicted using a pre-trained potential improvement prediction model, and the asset level improvement probability prediction value of the customer to be classified is obtained based on the inference and prediction results.

[0035] The asset level upgrade probability prediction value of this embodiment can be understood as the probability prediction value of the client's asset level upgrade. Research has found that the client's asset level upgrade is related to the client's various financial behaviors (such as the behavior of purchasing financial products). Such customers are generally more interested in various financial products. Therefore, this embodiment of the application uses the asset level upgrade probability prediction value to make an initial prediction of the client's asset marketing potential.

[0036] Step S130: performing an initial prediction of the asset marketing potential of the customer to be classified according to the asset grade promotion probability prediction value.

[0037] Step S140, when it is initially predicted that the customer to be classified has asset marketing potential, it is determined that the customer to be classified belongs to the first type of asset marketing customer group.

[0038] The asset marketing customer group in the embodiment of the present application, for example, includes a first type of asset marketing customer group and a second type of asset marketing customer group, wherein the first type of asset marketing customer group includes sub-customer groups corresponding to different asset levels, and the second type of asset marketing customer group also includes sub-customer groups corresponding to different asset potentials.

[0039] When it is determined that the customer to be classified has asset marketing potential according to the asset grade promotion probability prediction value, the customer to be classified can be added to the corresponding sub-customer group of the first type of asset marketing customer group.

[0040] Step 150, when it is initially predicted that the customer to be classified does not have asset marketing potential, second characteristic data related to asset potential is obtained based on the customer data, and a predicted value of the asset potential of the customer to be classified is obtained based on the second characteristic data, and the asset marketing potential of the customer to be classified is re-predicted based on the asset potential predicted value, and the asset marketing customer group to which the customer to be classified belongs is determined based on the re-prediction result.

[0041] The asset potential prediction value of the embodiment of the present application is used to predict the total value of the financial assets of the customer. The asset marketing potential of the customer is also related to the total value of the financial assets of the customer. Generally speaking, when the total value of the financial assets of the customer is high, the asset marketing potential of this type of customer is high, and when the total value of the financial assets of the customer is low, the asset marketing potential of this type of customer is low.

[0042] In actual applications, the total value of a customer's bank's financial assets may account for a relatively low proportion of the customer's total financial assets. Therefore, the embodiment of the present application predicts the total value of the financial assets of the customers to be classified through certain means. For customers who are initially predicted to have no asset marketing potential, the customer's asset marketing potential is further judged based on the customer's total financial assets to improve the accuracy of the prediction of the customer's asset marketing potential.

[0043] like Figure 1It can be seen from the bank customer classification method shown that the embodiment of the present application first uses the first characteristic data to obtain the asset grade improvement probability prediction value of the customer to be classified, and makes an initial prediction of the asset marketing potential of the customer to be classified based on the asset grade improvement probability prediction value. When it is initially predicted that the customer to be classified has asset marketing potential, it is determined that the customer to be classified belongs to the first type of asset marketing customer group; when it is initially predicted that the customer to be classified does not have asset marketing potential, the first characteristic data is used to obtain the asset potential prediction value of the customer to be classified, and the asset marketing potential of the customer to be classified is predicted again based on the asset potential prediction value, so as to determine the asset marketing customer group of the customer to be classified in combination with the asset grade improvement probability prediction value and the asset potential prediction value, thereby improving the accuracy of customer classification.

[0044] In some embodiments of the present application, the above step S120 acquires first characteristic data related to the asset level change according to the customer data, specifically including:

[0045] Acquire first target data related to asset level changes from the customer data of the customer to be classified through the expert experience database, wherein the first target data includes, for example, multi-dimensional data such as user basic information, contracted and held product information, asset information, transaction and account movement information, capital flow and RFM information;

[0046] By performing data preprocessing on the first target data, preprocessed feature data is obtained; wherein the data preprocessing includes:

[0047] Discrete feature variables are quantified, merged, deleted, etc., continuous feature variables are filled with the same type and outliers are removed, continuous feature variables are binned, and original variables are transformed to improve variable distribution or deviation; and to consider the diversity and interpretability of variables, variables are derived based on the original variables, such as by combining the original variables, reprocessing the original variables through different encoding methods, etc.

[0048] Perform correlation analysis on the preprocessed feature data and eliminate variables with low correlation with the target variable (asset grade improvement probability variable).

[0049] The first characteristic data having multi-dimensional customization is obtained by performing the above processing on the customer data.

[0050] In some embodiments of the present application, the step S120 above obtains the asset level promotion probability prediction value of the customer to be classified according to the first feature data, specifically including:

[0051] Obtaining a potential improvement prediction model, wherein the potential improvement prediction model includes potential improvement prediction sub-models corresponding to different asset levels, and the potential improvement prediction sub-models are pre-trained models constructed based on the Logistic algorithm;

[0052] Acquire the asset level of the customer to be classified according to the customer data;

[0053] Determine a target potential improvement prediction sub-model according to the asset level of the customer to be classified, wherein the benchmark asset level of the target potential improvement prediction sub-model corresponds to the asset level of the customer to be classified;

[0054] The target potential improvement prediction sub-model is used to perform inference prediction on the first feature data to obtain a predicted value of the asset level improvement probability of the customer to be classified.

[0055] The embodiment of the present application can construct asset grade information in advance, and the asset grade information includes asset grade measurement indicators and asset grade division situations. For example, asset grade information can be divided according to the average daily AUM (Asset Under Management) value. The specific division situation can be set according to actual business needs. The embodiment of the present application does not impose specific restrictions on this. In addition, other asset indicators can also be used as asset grade measurement indicators. Technical personnel in this field can reasonably select asset grade measurement indicators according to business needs.

[0056] Accordingly, this embodiment obtains the daily average AUM value of the customer to be classified based on the above customer data, and then determines the asset grade of the customer to be classified based on the daily average AUM value of the customer to be classified and the asset grade information constructed in advance.

[0057] Usually, customer behaviors of different asset levels are significantly different. In order to accurately predict customer behavior preferences, the asset level can be set according to the average daily AUM value of a preset time period (for example, the past month). For example, the three asset levels A, B, and C are included. Data sets for the three asset levels are constructed separately, and the same set of Logistic algorithm models are trained using the three data sets to obtain the potential improvement prediction sub-model corresponding to each asset level.

[0058] When constructing the data set corresponding to each asset level, the asset enhancement target value corresponding to each asset level is set respectively, and the customer samples whose total financial assets of the bank customers reach the set asset enhancement target value are taken as positive samples. Conversely, the customer samples whose total financial assets of the bank customers do not reach the set asset enhancement target value are taken as negative samples. Then, the customer data of the positive and negative samples are processed to obtain the first characteristic data of the positive and negative samples, thereby completing the construction of the data set.

[0059] In the model training stage, multiple indicators such as ROC curve, Recall value, Precision value and Lift value can be used to evaluate the potential improvement prediction sub-model, and the model variables and parameters are continuously adjusted until each indicator achieves the expected effect. Of course, in other embodiments of the present application, other algorithms can also be used to construct the initial model architecture and other model evaluation indicators can be used, and the embodiments of the present application do not specifically limit this.

[0060] In some embodiments of the present application, the initial prediction of the asset marketing potential of the customer to be classified according to the asset grade promotion probability prediction value in the above step S130 specifically includes:

[0061] Comparing the predicted value of the probability of asset grade improvement of the customer to be classified with the corresponding threshold value of the probability of asset grade improvement;

[0062] When the predicted value of the probability of asset grade improvement of the customer to be classified is greater than the threshold value of the probability of asset grade improvement, it is initially predicted that the customer to be classified has asset marketing potential. At this time, it can be determined that the customer to be classified belongs to the first type of asset marketing customer group, and the sub-customer group of the first type of asset marketing customer group to which the customer to be classified belongs is determined according to the asset grade of the customer to be classified. For example, when the asset grade of the customer to be classified is A, the customer to be classified is classified into the A-level sub-customer group of the first type of asset marketing customer group;

[0063] When the predicted value of the probability of asset level improvement of the customer to be classified is not greater than the threshold value of the probability of asset level improvement, it is initially predicted that the customer to be classified does not have asset marketing potential, and it should be determined whether the predicted value of the asset potential of the customer to be classified meets the standard. If it meets the standard, it means that the customer to be classified has a high asset potential. At this time, the asset marketing customer group to which the customer to be classified belongs should be determined based on the predicted value of the asset potential. If it does not meet the standard, it means that the customer to be classified has a low asset potential. At this time, it can be determined that the customer to be classified has no asset marketing potential and does not belong to the various asset marketing customer groups pre-set in the embodiment of the present application.

[0064] In this embodiment, the same or different asset grade upgrade probability thresholds can be designed for different asset grades, that is, the asset grades ABC mentioned above can have the same or different asset grade upgrade probability thresholds. Those skilled in the art can set the asset grade upgrade probability thresholds based on actual application conditions and with the help of experience.

[0065] In some embodiments of the present application, obtaining second characteristic data related to asset potential according to the customer data in the above step S150 specifically includes:

[0066] Acquire second target data related to asset potential from the customer data of the customer to be classified through the expert experience database, wherein the second target data includes, for example, multi-dimensional data such as user basic information, contracted and held product information, medium- and long-term transactions, and asset information;

[0067] The second target data is subjected to data standardization processing, and the second feature data is obtained according to the standardization processing result, wherein the second feature data is multi-dimensional feature data.

[0068] In practical applications, customer data of a preset time window (eg, one year) may be extracted from a data warehouse and processed to obtain second feature data in the form of a feature wide table.

[0069] In some embodiments of the present application, obtaining the asset potential prediction value of the customer to be classified according to the second feature data in the above step S150 specifically includes:

[0070] Acquire a pre-built reference customer group, the reference customer group including a plurality of reference customers, the asset levels of the plurality of reference customers being greater than a preset asset level and including various asset levels above the predicted asset level;

[0071] Acquire feature similarity information between the customer to be classified and the reference customer group based on the second feature data of the customer to be classified and the second feature data of the reference customer;

[0072] The asset potential prediction value of the customer to be classified is obtained according to the feature similarity information.

[0073] As mentioned above, in general, the higher the asset level, the higher the asset marketing potential. Therefore, this embodiment uses reference customers whose asset levels are greater than the preset asset level to construct a reference customer group, where the preset asset level can be set based on experience. When a customer's asset level is greater than the preset asset level, it can be determined that the customer has asset marketing potential. Considering that the bank's customers are large in scale and widely distributed, and the behavioral preferences of customers with different asset levels are significantly different, this embodiment makes the asset levels of all reference customers in the reference customer group include various asset levels above the predicted asset level.

[0074] Considering that customers of the same asset level have similar behavioral preferences, for example, they have certain similarities in terms of residential address, travel, and financial behavior. Based on this, the embodiment of the present application pre-constructs the reference customer group, screens target customers with similar characteristics to the customers to be classified from the reference customer group through the second characteristic data, and uses the target customers' asset information to predict the asset potential prediction value of the customers to be classified. In this way, even if the total value of the bank's financial assets of the customers to be classified is low, then according to the asset potential prediction method of this embodiment, the "real" total value of the financial assets of the customers to be classified can be predicted to a certain extent, avoiding the mistaken classification of customers with low total value of financial assets of the bank as customers without asset marketing potential.

[0075] In some possible implementations of this embodiment, the second feature data includes a plurality of feature variables, and obtaining feature similarity information between the customer to be classified and the reference customer according to the second feature data of the customer to be classified and the second feature data of the reference customer specifically includes:

[0076] Obtaining a characteristic distance between each characteristic variable of the second characteristic data of the customer to be classified and a corresponding characteristic variable of the second characteristic data of the reference customer, wherein the characteristic distance includes, for example, a Euclidean distance;

[0077] The feature similarity information between the customer to be classified and the reference customer is obtained by weighted summing the feature distances of the multiple feature variables.

[0078] Accordingly, obtaining the asset potential prediction value of the customer to be classified according to the feature similarity information specifically includes:

[0079] Comparing the feature similarity information with a preset feature similarity threshold;

[0080] When the maximum value of the feature similarity information is not greater than the preset feature similarity threshold, it is determined that the asset potential prediction value of the customer to be classified does not meet the standard. At this time, it can be determined that the customer to be classified does not belong to the various asset marketing customer groups preset in the embodiment of the present application;

[0081] When the maximum value of the feature similarity information is greater than the preset feature similarity threshold, it is determined that the asset potential prediction value of the customer to be classified meets the standard, and the target customer is obtained from the reference customer group based on the feature similarity information, and the asset potential prediction value is obtained based on the asset information of the target customer and the asset information of the customer to be classified.

[0082] For example, the feature similarity information between the customer I to be classified and M (M is a natural number greater than 100) reference customers in the reference customer group is calculated, where the feature similarity information of more than K (optionally, K is a natural number not greater than 10) reference customers is greater than a preset feature similarity threshold, then the K reference customers with the highest feature similarity information are selected as target customers, and the average of the average daily AUM values ​​of the K target customers last month is calculated, and the average is used as the asset potential prediction value of the customer I to be classified.

[0083] In some embodiments of the present application, the above step S150 re-predicts the asset marketing potential of the customer to be classified according to the asset potential prediction value, and determines the asset marketing customer group to which the customer to be classified belongs according to the re-prediction result, specifically including:

[0084] When the asset potential prediction value does not meet the standard, determining that the customer to be classified does not have asset marketing potential;

[0085] When the asset potential prediction value reaches the standard, it is determined that the customer to be classified has asset marketing potential and belongs to the second type of asset marketing customer group.

[0086] In some embodiments of the present application, when it is determined that the customer to be classified belongs to the second type of asset marketing customer group, Figure 1 The bank customer classification methods in the 2014 Bank of China also include:

[0087] Acquire asset information and fund entry history information of the customer to be classified according to the customer data, wherein the fund entry history information includes large-amount fund entry within a set time window and no large-amount fund entry within a set time window, wherein the large-amount fund here can be set according to demand, for example, a fund amount of more than 50,000 is determined as a large-amount fund;

[0088] According to the asset potential forecast value and the asset information of the customer to be classified, obtaining the asset proportion of the customer to be classified in the bank;

[0089] The sub-customer group of the second type of asset marketing customer group to which the customer to be classified belongs is determined based on the proportion of the bank's assets and the historical information on funds deposited into the account of the customer to be classified.

[0090] In some possible implementations of this embodiment, the asset information includes a daily average AUM value, and obtaining the asset proportion of the customer to be classified in the bank according to the asset potential forecast value and the asset information of the customer to be classified specifically includes:

[0091] The ratio of the daily average AUM value of the customer to be classified to the asset potential forecast value is used as the asset proportion of the customer to be classified in the bank.

[0092] It should be understood that when the asset information is other indicator values, the asset ratio of the bank can also be calculated for other indicators using a similar calculation method, which is not explained one by one in this embodiment.

[0093] In some other possible implementations of this embodiment, determining the sub-customer group of the second type of asset marketing customer group to which the customer to be classified belongs according to the bank's asset ratio of the customer to be classified and the fund entry history information specifically includes:

[0094] When the fund inflow history information includes a large amount of funds inflow within a set time window, determining that the customer to be classified belongs to a historical high inflow sub-customer group corresponding to the bank's capital ratio according to the bank's capital ratio;

[0095] When the fund deposit history information includes that no large amount of funds have been deposited within a set time window, it is determined that the customer to be classified belongs to a non-historical high-inflow sub-customer group corresponding to the bank's capital ratio based on the bank's capital ratio.

[0096] For example, when a large amount of funds are identified to be deposited into the account of customer I to be classified based on the identifiable characteristics in the same quarter within two years, customer I to be classified will be classified into a sub-customer group with historical high inflows. An asset marketing strategy can be designed for the sub-customer group with historical high inflows, so as to carry out asset marketing for the sub-customer group with historical high inflows in the same quarter of the current year.

[0097] Combine the following Figure 2 and Figure 3 The bank customer classification method of the embodiment of the present application is described in detail.

[0098] like Figure 2 As shown, the computing device of the embodiment of the present application may further include a model calculation module, a marketing strategy association module and a data push module. The model calculation module is used to execute the bank customer classification method in the above embodiment to obtain a customer list of each asset marketing customer group; the marketing strategy association module associates the corresponding asset marketing strategy for each asset marketing customer group. The asset marketing strategy here can be the latest marketing strategy uploaded to the computing device by the asset marketing management platform. Optionally, the asset marketing strategy includes a strategy combination such as product recommendation and marketing method; the data push module is used to push the asset marketing strategy of each asset marketing customer group to the operating personnel, for example, to the bank's internal management system for the convenience of the operating personnel to view.

[0099] Among them, Figure 3As shown, the model calculation module first constructs asset grade ABC according to the statistical mean of the daily average AUM value of bank customers in the past month, and uses the customer data of bank customers in the past month to construct three data sets for asset grade ABC for training the Logistics algorithm model, obtains three potential improvement prediction sub-models corresponding to asset grade ABC, and determines the target potential improvement prediction sub-model according to the asset grade of the customer to be classified to predict the probability of asset grade improvement of the customer to be classified in the next month, and obtains the asset grade improvement probability prediction value y, and compares the asset grade improvement probability prediction value y with the asset grade improvement probability threshold a. When y>a, it means that the customer to be classified is a high-upgrade potential customer. At this time, it can be determined that the customer to be classified belongs to the first type of asset marketing customer group. If the asset grade of the customer to be classified is A, it is further determined that the customer to be classified belongs to the A-level sub-customer group of the first type of asset marketing customer group.

[0100] When y≤a, it means that the customer to be classified is a customer with low upgrade potential. At this time, K target customers who are most similar to the customer to be classified are screened out from the customer group with higher asset level by referring to the customer group. The average of the daily average AUM values ​​of the K target customers is used as the asset potential prediction value of the customer to be classified. The bank's asset ratio and funds receipt history information are calculated to classify customers with low upgrade potential and low asset ratio of the bank into asset marketing potential.

[0101] The present application also provides a bank customer classification device, such as Figure 4 As shown, a schematic diagram of the structure of a bank customer classification device in an embodiment of the present application is provided, wherein the bank customer classification device 400 comprises: a data acquisition unit 410, a data processing unit 420 and a prediction and classification unit 430, wherein:

[0102] A data acquisition unit 410 is used to acquire customer data of customers to be classified;

[0103] The data processing unit 420 is used to obtain first characteristic data related to asset grade change according to the customer data, and obtain a predicted value of asset grade improvement probability of the customer to be classified according to the first characteristic data;

[0104] The prediction and classification unit 430 is used to make an initial prediction of the asset marketing potential of the customer to be classified based on the predicted value of the asset grade upgrade probability, and when it is initially predicted that the customer to be classified has the asset marketing potential, determine that the customer to be classified belongs to the first type of asset marketing customer group; and when it is initially predicted that the customer to be classified does not have the asset marketing potential, obtain second characteristic data related to the asset potential based on the customer data, and obtain the asset potential prediction value of the customer to be classified based on the second characteristic data, re-predict the asset marketing potential of the customer to be classified based on the asset potential prediction value, and determine the asset marketing customer group to which the customer to be classified belongs based on the re-prediction result.

[0105] In some embodiments of the present application, the data processing unit 420 is specifically used to obtain a potential improvement prediction model, the potential improvement prediction model includes potential improvement prediction sub-models corresponding to different asset levels, and the potential improvement prediction sub-model is a pre-trained model constructed based on the Logistic algorithm; the asset level of the customer to be classified is obtained according to the customer data; the target potential improvement prediction sub-model is determined according to the asset level of the customer to be classified, and the benchmark asset level of the target potential improvement prediction sub-model corresponds to the asset level of the customer to be classified; and the first feature data is inferred and predicted using the target potential improvement prediction sub-model to obtain a predicted value of the asset level improvement probability of the customer to be classified.

[0106] In some embodiments of the present application, the prediction and classification unit 430 is specifically used to compare the predicted value of the probability of asset level improvement of the customer to be classified with the corresponding asset level improvement probability threshold; when the predicted value of the probability of asset level improvement of the customer to be classified is greater than the asset level improvement probability threshold, it is initially predicted that the customer to be classified has asset marketing potential; when the predicted value of the probability of asset level improvement of the customer to be classified is not greater than the asset level improvement probability threshold, it is initially predicted that the customer to be classified does not have asset marketing potential.

[0107] In some embodiments of the present application, the asset potential prediction value includes meeting the standard and not meeting the standard. The prediction and classification unit 430 is specifically used to determine that the customer to be classified does not have asset marketing potential when the asset potential prediction value does not meet the standard; when the asset potential prediction value meets the standard, determine that the customer to be classified has asset marketing potential and belongs to the second category of asset marketing customers.

[0108] In some embodiments of the present application, the prediction and classification unit 430 is also used to obtain the asset information and fund deposit history information of the customer to be classified based on the customer data when it is determined that the customer to be classified belongs to the second type of asset marketing customer group, the fund deposit history information including large fund deposits within a set time window and no large fund deposits within a set time window; obtain the asset ratio of the customer to be classified in the bank based on the asset potential prediction value and the asset information of the customer to be classified; determine the sub-customer group of the second type of asset marketing customer group to which the customer to be classified belongs based on the asset ratio of the customer to be classified in the bank and the fund deposit history information.

[0109] In some embodiments of the present application, the prediction and classification unit 430 is specifically used to obtain a pre-constructed reference customer group, wherein the reference customer group includes multiple reference customers, and the asset level of the multiple reference customers is greater than the preset asset level and includes various asset levels above the predicted asset level; obtain characteristic similarity information between the customer to be classified and the reference customer group based on the second characteristic data of the customer to be classified and the second characteristic data of the reference customer; and obtain the asset potential prediction value of the customer to be classified based on the characteristic similarity information.

[0110] In some embodiments of the present application, the prediction and classification unit 430 is specifically used to compare the feature similarity information with a preset feature similarity threshold; when the maximum value of the feature similarity information is not greater than the preset feature similarity threshold, it is determined that the asset potential prediction value of the customer to be classified does not meet the standard; when the maximum value of the feature similarity information is greater than the preset feature similarity threshold, it is determined that the asset potential prediction value of the customer to be classified meets the standard, and the target customer is obtained from the reference customer group based on the feature similarity information, and the asset potential prediction value is obtained based on the asset information of the target customer and the asset information of the customer to be classified.

[0111] It can be understood that the above-mentioned bank customer classification can implement each step of the bank customer classification method provided in the aforementioned embodiment, and the relevant explanations about the bank customer classification method are applicable to the bank customer classification device, which will not be repeated here.

[0112] Figure 5 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 5 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0113] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0114] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0115] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a bank customer classification device at the logical level. The processor executes the program stored in the memory and is specifically used to perform the following operations:

[0116] Obtain customer data of the customer to be classified;

[0117] Acquire first characteristic data related to asset grade change according to the customer data, and acquire a predicted value of the probability of asset grade improvement of the customer to be classified according to the first characteristic data;

[0118] Performing an initial forecast of the asset marketing potential of the customer to be classified according to the asset grade improvement probability forecast value;

[0119] When initially predicting that the customer to be classified has asset marketing potential, determining that the customer to be classified belongs to the first type of asset marketing customer group;

[0120] When it is initially predicted that the customer to be classified does not have asset marketing potential, second characteristic data related to asset potential is obtained based on the customer data, and a predicted value of the asset potential of the customer to be classified is obtained based on the second characteristic data, and the asset marketing potential of the customer to be classified is re-predicted based on the predicted value of asset potential, and the asset marketing customer group to which the customer to be classified belongs is determined based on the re-prediction result.

[0121] The above application Figure 1The method performed by the bank customer classification device disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by a hardware integrated logic circuit in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above-mentioned bank customer classification method in combination with its hardware.

[0122] The electronic device may also perform Figure 1 The method is implemented by the bank customer classification device in Figure 1 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0123] The present application also provides a computer-readable storage medium, which stores one or more programs, wherein the one or more programs include instructions, which, when executed by an electronic device including multiple application programs, enable the electronic device to execute Figure 1 The method performed by the bank customer classification device in the illustrated embodiment is specifically used to perform:

[0124] Obtain customer data of the customer to be classified;

[0125] Acquire first characteristic data related to asset grade change according to the customer data, and acquire a predicted value of the probability of asset grade improvement of the customer to be classified according to the first characteristic data;

[0126] Performing an initial forecast of the asset marketing potential of the customer to be classified according to the asset grade improvement probability forecast value;

[0127] When initially predicting that the customer to be classified has asset marketing potential, determining that the customer to be classified belongs to the first type of asset marketing customer group;

[0128] When it is initially predicted that the customer to be classified does not have asset marketing potential, second characteristic data related to asset potential is obtained based on the customer data, and a predicted value of the asset potential of the customer to be classified is obtained based on the second characteristic data, and the asset marketing potential of the customer to be classified is re-predicted based on the predicted value of asset potential, and the asset marketing customer group to which the customer to be classified belongs is determined based on the re-prediction result.

[0129] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0130] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0131] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0132] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0133] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0134] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0135] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0136] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0137] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0138] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A bank customer classification method, characterized in that: The bank customer classification method comprises: Obtain customer data of the customer to be classified; Acquire first characteristic data related to asset grade change according to the customer data, and acquire a predicted value of the probability of asset grade improvement of the customer to be classified according to the first characteristic data; Performing an initial forecast of the asset marketing potential of the customer to be classified according to the asset grade upgrade probability forecast value; When initially predicting that the customer to be classified has asset marketing potential, determining that the customer to be classified belongs to the first type of asset marketing customer group; When it is initially predicted that the customer to be classified does not have asset marketing potential, second characteristic data related to asset potential is obtained based on the customer data, and a predicted value of the asset potential of the customer to be classified is obtained based on the second characteristic data, and the asset marketing potential of the customer to be classified is re-predicted based on the predicted value of asset potential, and the asset marketing customer group to which the customer to be classified belongs is determined based on the re-prediction result.

2. The bank customer classification method according to claim 1, characterized in that: The step of obtaining the predicted value of the probability of asset level improvement of the customer to be classified according to the first characteristic data includes: Obtaining a potential improvement prediction model, wherein the potential improvement prediction model includes potential improvement prediction sub-models corresponding to different asset levels, and the potential improvement prediction sub-models are pre-trained models constructed based on the Logistic algorithm; Acquire the asset level of the customer to be classified according to the customer data; Determine a target potential improvement prediction sub-model according to the asset level of the customer to be classified, wherein the benchmark asset level of the target potential improvement prediction sub-model corresponds to the asset level of the customer to be classified; The target potential improvement prediction sub-model is used to perform inference prediction on the first feature data to obtain a predicted value of the asset level improvement probability of the customer to be classified.

3. The bank customer classification method according to claim 2, characterized in that: The initial prediction of the asset marketing potential of the customer to be classified according to the asset grade promotion probability prediction value includes: Comparing the predicted value of the probability of asset grade improvement of the customer to be classified with the corresponding threshold value of the probability of asset grade improvement; When the predicted value of the probability of asset grade improvement of the customer to be classified is greater than the probability threshold of asset grade improvement, it is initially predicted that the customer to be classified has asset marketing potential; When the predicted value of the probability of asset grade improvement of the customer to be classified is not greater than the threshold value of the probability of asset grade improvement, it is initially predicted that the customer to be classified has no asset marketing potential.

4. The bank customer classification method according to claim 1, characterized in that: The asset potential prediction value includes reaching the standard and not reaching the standard, and re-predicting the asset marketing potential of the customer to be classified according to the asset potential prediction value, and determining the asset marketing customer group to which the customer to be classified belongs according to the re-prediction result, includes: When the asset potential prediction value does not meet the standard, determining that the customer to be classified does not have asset marketing potential; When the asset potential prediction value reaches the standard, it is determined that the customer to be classified has asset marketing potential and belongs to the second type of asset marketing customer group.

5. The bank customer classification method according to claim 4, characterized in that: When it is determined that the customer to be classified belongs to the second type of asset marketing customer group, the method further includes: Acquire asset information and fund entry history information of the customer to be classified according to the customer data, wherein the fund entry history information includes whether there is a large amount of fund entry within a set time window and whether there is no large amount of fund entry within the set time window; According to the asset potential forecast value and the asset information of the customer to be classified, obtaining the asset proportion of the customer to be classified in the bank; The sub-customer group of the second type of asset marketing customer group to which the customer to be classified belongs is determined based on the proportion of the bank's assets and the historical information on funds deposited into the account of the customer to be classified.

6. The bank customer classification method according to claim 1, characterized in that: The step of obtaining the asset potential prediction value of the customer to be classified according to the second characteristic data includes: Acquire a pre-built reference customer group, the reference customer group including a plurality of reference customers, the asset levels of the plurality of reference customers being greater than a preset asset level and including various asset levels above the predicted asset level; Acquire feature similarity information between the customer to be classified and the reference customer group based on the second feature data of the customer to be classified and the second feature data of the reference customer; The asset potential prediction value of the customer to be classified is obtained according to the feature similarity information.

7. The bank customer classification method according to claim 6, characterized in that: The step of obtaining the asset potential prediction value of the customer to be classified according to the feature similarity information includes: Comparing the feature similarity information with a preset feature similarity threshold; When the maximum value of the feature similarity information is not greater than the preset feature similarity threshold, it is determined that the asset potential prediction value of the customer to be classified does not meet the standard; When the maximum value of the feature similarity information is greater than the preset feature similarity threshold, it is determined that the asset potential prediction value of the customer to be classified meets the standard, and the target customer is obtained from the reference customer group based on the feature similarity information, and the asset potential prediction value is obtained based on the asset information of the target customer and the asset information of the customer to be classified.

8. A bank customer classification device, characterized in that: The bank customer classification device comprises: A data acquisition unit, used to acquire customer data of customers to be classified; A data processing unit, configured to obtain first characteristic data related to asset grade change according to the customer data, and obtain a predicted value of the probability of asset grade improvement of the customer to be classified according to the first characteristic data; A prediction and classification unit is used to make an initial prediction of the asset marketing potential of the customer to be classified according to the predicted value of the probability of asset grade improvement, and when it is initially predicted that the customer to be classified has the asset marketing potential, determine that the customer to be classified belongs to the first type of asset marketing customer group; and when it is initially predicted that the customer to be classified does not have the asset marketing potential, obtain second characteristic data related to the asset potential according to the customer data, and obtain the asset potential prediction value of the customer to be classified according to the second characteristic data, re-predict the asset marketing potential of the customer to be classified according to the asset potential prediction value, and determine the asset marketing customer group to which the customer to be classified belongs according to the re-prediction result.

9. An electronic device, comprising: processor; as well as A memory arranged to store computer executable instructions, which, when executed, cause the processor to execute the bank customer classification method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, enables the electronic device to execute the bank customer classification method according to any one of claims 1 to 7.

Images