Data processing method, electronic device, storage medium and program product

Abstract

The embodiment of the invention provides a data processing method, electronic equipment, a storage medium and a program product. The method comprises the following steps: judging which type a target customer to be processed belongs to; according to the type, feature data related to the type of customers is found; and according to the feature data, determining a link in which the target customer is located, and pushing corresponding activity information to the target customer according to an activity pushing strategy of the corresponding link. When the feature data of the target customer is determined, not only are behavior features of the credit card application program considered, but also card use features of the credit card and card use and behavior composite features are considered, so that the determined features can be more comprehensive, and activity information can be pushed to the target customer more accurately.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a data processing method, electronic device, storage medium, and program product. Background Art

[0002] In today's digital age, with the rapid development of mobile internet, customer engagement for various applications and credit cards has become increasingly important. Customers' functional demands for applications are becoming increasingly diverse, with customers expecting not only convenient payment, social interaction, and entertainment experiences, but also personalized service recommendations.

[0003] Therefore, how to accurately push event information to customers and provide high-quality services is an urgent problem that needs to be solved today. Summary of the Invention

[0004] The embodiments of the present application provide a data processing method, an electronic device, a storage medium, and a program product for accurately pushing activity information to customers.

[0005] In a first aspect, an embodiment of the present application provides a data processing method, the method comprising:

[0006] Determine the type of target customers to be processed, where the target customer type is at least one of the following: customers who have not linked their credit card application to a credit card, customers who have active credit card application-linked cards, customers with a churn warning for credit card application, or customers who have churned from credit card application;

[0007] Determine corresponding characteristic data based on the type of target customer, the characteristic data including at least one of the following: card usage characteristics, credit card application behavior characteristics, and card usage and behavior composite characteristics, wherein the card usage and behavior composite characteristics are determined based on the card usage characteristics and credit card application behavior characteristics;

[0008] Determine the activity push strategy for the corresponding link based on the characteristic data, and the corresponding link is one of the following: credit card application card binding link, credit card application active customer maintenance link, credit card application customer churn warning link, and credit card application churn customer recall link;

[0009] Push corresponding activity information to target client terminals according to the corresponding activity push strategy.

[0010] In one possible implementation, determining the type of target customer to be processed includes:

[0011] The type of the target customer is determined based on at least one of whether the target customer has bound a corresponding credit card in the credit card application and whether the target customer has started the credit card application within a preset time period.

[0012] In a possible implementation, determining corresponding characteristic data according to the type of target customer includes:

[0013] If it is determined that the target customer is a customer who has not bound a card to a credit card application, then the corresponding feature data is determined to be the first card usage feature;

[0014] If the target customer is determined to be an active customer who has bound a credit card to an application, the corresponding feature data is determined to be the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature;

[0015] If it is determined that the target customer is a credit card application churn warning customer, the corresponding feature data is determined to be the third card usage feature, the second application behavior feature, and the second card usage and behavior composite feature;

[0016] If it is determined that the type of target customer is a credit card application churn customer, the corresponding feature data is determined to be the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature.

[0017] In a possible implementation, the target customer type is a customer who has not linked a credit card application to a credit card application. The activity push strategy for the corresponding link is determined based on the characteristic data, including:

[0018] Based on the characteristic data, determine whether the target customer is a recommended customer in the credit card application binding process;

[0019] In response to the target customer being a push customer for a credit card application binding phase, an activity push strategy for the credit card application binding phase is determined.

[0020] In one possible implementation, determining whether the target customer is a push customer in the credit card application binding process based on the characteristic data includes:

[0021] Inputting the first card usage feature into a card binding propensity prediction model, training the card binding propensity prediction model until convergence, and using the card binding propensity prediction model that has been trained to convergence to determine a card binding propensity score corresponding to the target customer in the card binding process of the credit card application. The card binding propensity prediction model that has been trained to convergence is located in a card binding module of the credit card application.

[0022] Based on the card binding propensity score, determine whether the target customer is a recommended customer for the credit card application card binding process.

[0023] In a possible implementation, before inputting the first card usage feature into the card binding tendency prediction model, the method further includes:

[0024] Based on the churn customer recall module, obtain the churn recall scores corresponding to the credit card application churn customers among the target customers;

[0025] Based on the churn recall score, determine the non-churn recall intention customers and delete them from the target customers.

[0026] In one possible implementation, determining whether a target customer is a recommended customer for a credit card application program to bind a card based on a card binding propensity score includes:

[0027] Determine the card binding tendency level based on the card binding tendency score;

[0028] In response to the card binding tendency level being the preset information push level, determining the target customer as a push customer for the credit card application program in the card binding process;

[0029] In response to the card binding tendency level not being the preset information push level, it is determined that the target customer is not a push customer in the card binding link of the credit card application.

[0030] In one possible implementation, determining an activity push strategy for the card application binding phase includes:

[0031] Determine the corresponding promotion channel based on the target customers' card binding tendency level;

[0032] Obtain customer preference tags based on active customer maintenance in credit card applications;

[0033] Based on the corresponding activity push channels and customer preference tags, determine the activity push strategy for the corresponding card application binding process.

[0034] Push corresponding activity information to target customer terminals according to the corresponding activity push strategy, including:

[0035] Push activity information that matches the target customer's preference tags to the target customer terminal based on the corresponding activity push channel.

[0036] In one possible implementation, the target customer type is an active customer who has bound their credit card to an application. The activity push strategy for the corresponding link is determined based on the characteristic data, including:

[0037] Determining at least one of a target customer's credit card application active behavior tag and a customer preference tag based on the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature;

[0038] Based on the credit card application binding process, obtain the target customer's activity push channel;

[0039] Determine a corresponding activity push strategy based on at least one of the target customer's credit card application active behavior tag, the customer preference tag, and / or the activity push channel;

[0040] Push corresponding activity information to target customer terminals according to the corresponding activity push strategy, including:

[0041] Pushing activity information to the target customer terminal via the credit card application based on at least one of the target customer's credit card application active behavior tag and customer preference tag;

[0042] And / or, at least one of the target customer's credit card application active behavior tag and the customer preference tag is used to push activity information to the target customer terminal through the target customer's activity push channel.

[0043] In a possible implementation, the target customer type is a credit card application churn warning customer, and the activity push strategy for the corresponding link is determined based on the characteristic data, including:

[0044] Based on the active customer maintenance phase of the credit card application, obtain the active behavior label of the credit card application;

[0045] Inputting the third card usage feature, the second application behavior feature, the second card usage and behavior composite feature, and the credit card application active behavior label into a churn warning model, training the churn warning model until convergence, and using the churn warning model that has been trained to convergence to determine a churn warning score corresponding to the target customer in the credit card application customer churn warning phase. The churn warning model that has been trained to convergence is located in a customer churn warning module.

[0046] Based on the third card usage feature, obtain the target customer's card churn category label. The target customer's card usage flow type label is used to indicate the probability that the target customer will not use the corresponding credit card;

[0047] Determine the churn customer group to which the target customer belongs based on their churn warning score and card usage churn category label in the credit card application's customer churn warning process;

[0048] Based on the churn customer group to which the target customer belongs, determine the corresponding activity push strategy in the customer churn warning link of the credit card application.

[0049] In one possible implementation, determining an activity push strategy corresponding to the credit card application customer churn warning phase based on the churn customer group to which the target customer belongs includes:

[0050] If it is determined that the target customer belongs to the first-level churn customer group, the activity push strategy corresponding to the customer churn warning phase of the credit card application is determined to be the first-level churn push strategy;

[0051] If it is determined that the target customer belongs to the second-level churn customer group, the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the second-level churn push strategy;

[0052] If it is determined that the target customer belongs to the third-level churn customer group, then the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the third-level churn push strategy;

[0053] If it is determined that the target customer belongs to the fourth-level churn customer group, then the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the fourth-level churn push strategy;

[0054] In one possible implementation, determining the churn customer group to which the target customer belongs based on the churn warning score and card churn category label corresponding to the target customer in the customer churn warning phase of the credit card application includes:

[0055] If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to the first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to the second threshold, then the target customer belongs to the first-level churn customer group;

[0056] If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to the first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than the second threshold, then the target customer belongs to the second-level churn customer group;

[0057] If the target customer's churn warning score in the credit card application's customer churn warning phase is less than the first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to the second threshold, then the target customer belongs to the third-level churn customer group;

[0058] If the target customer's churn warning score corresponding to the credit card application's customer churn warning link is less than the first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than the second threshold, then the target customer belongs to the fourth-level churn customer group.

[0059] In a possible implementation, the target customer type is a credit card application churn customer, and the activity push strategy for the corresponding link is determined based on the characteristic data, including:

[0060] inputting the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature into a churn recall model, training the churn recall model until convergence, and using the churn recall model that has been trained to convergence to determine a churn recall score corresponding to the target customer in the credit card application churn customer recall phase, wherein the churn recall model that has been trained to convergence is located in a churn customer recall module;

[0061] Obtain the target customer's churn warning score from the customer churn warning module;

[0062] Determine the recall customer group to which the target customer belongs based on their churn recall score and churn warning score in the credit card application's churn recall phase;

[0063] Obtaining the target customer's card transaction preference tag based on the fourth card usage characteristic;

[0064] Based on at least one of the recall customer group to which the target customer belongs and the card transaction preference tag, a corresponding activity push strategy for recalling churned customers in a credit card application is determined.

[0065] In one possible implementation, determining the recall customer group to which the target customer belongs based on the target customer's churn recall score and churn warning score corresponding to the churn recall phase of the credit card application includes:

[0066] If the target customer's corresponding churn recall score is greater than or equal to the third threshold, and the churn warning score is greater than or equal to the fourth threshold, the target customer's recall customer group is determined to be the first-level recall customer group;

[0067] If the churn recall score corresponding to the target customer is greater than or equal to the third threshold, and the churn warning score is less than the fourth threshold, the target customer is determined to belong to the second-level recall customer group;

[0068] If the churn recall score corresponding to the target customer is less than the third threshold, and the churn recall score corresponding to the target customer is greater than or equal to the fifth threshold, the target customer is determined to belong to the third-level recall customer group;

[0069] If the churn recall score corresponding to the target customer is less than the fifth threshold, the target customer is determined to belong to the fourth-level recall customer group;

[0070] In one possible implementation, determining a push strategy for an activity in the credit card application program to recall churned customers based on at least one of the target customer's recall customer group and the card transaction preference tag includes:

[0071] If it is determined that the churned customer group to which the target customer belongs is a first-level recall customer group, then the activity push strategy corresponding to the churned customer recall phase of the credit card application is determined to be a first-level recall push strategy, which includes pushing activity information to the target customer;

[0072] If it is determined that the target customer belongs to a second-level recall customer group, then the corresponding activity push strategy for the lapsed customer recall phase of the credit card application is determined to be a second-level recall push strategy. The second-level push strategy includes pushing activity information matching the card transaction preference tag to the target customer.

[0073] If it is determined that the target customer belongs to the third-level recall customer group, the corresponding activity push strategy for the churn customer recall phase of the credit card application is determined to be the third-level recall push strategy. The third-level recall push strategy includes pushing activity information matching the card transaction preference tag to the target customer.

[0074] If it is determined that the churned customer group to which the target customer belongs is the fourth-level recall customer group, then the activity push strategy corresponding to the churned customer recall link of the credit card application is determined to be the fourth-level recall push strategy.

[0075] In a second aspect, an embodiment of the present application provides a data processing device, the device comprising:

[0076] a determination module, configured to determine a type of target customer to be processed, the type of target customer being at least one of the following: a customer who has not bound a credit card application, an active customer who has bound a credit card application, a customer who has been warned of churn in a credit card application, or a customer who has churned in a credit card application;

[0077] The determination module is further configured to determine corresponding characteristic data based on the type of target customer, the characteristic data including at least one of the following: card usage characteristics, credit card application behavior characteristics, and card usage and behavior composite characteristics, wherein the card usage and behavior composite characteristics are determined based on the card usage characteristics and the credit card application behavior characteristics;

[0078] A processing module is used to determine an activity push strategy for a corresponding link based on the characteristic data, where the corresponding link is one of the following: a credit card application binding link, a credit card application active customer maintenance link, a credit card application customer churn warning link, and a credit card application churn customer recall link;

[0079] The push module is used to push corresponding activity information to the target client terminal according to the corresponding activity push strategy.

[0080] In a possible implementation, the determination module is specifically configured to:

[0081] The type of the target customer is determined based on at least one of whether the target customer has bound a corresponding credit card in the credit card application and whether the target customer has started the credit card application within a preset time period.

[0082] In a possible implementation, the determination module is specifically configured to:

[0083] If it is determined that the target customer is a customer who has not bound a card to a credit card application, then the corresponding feature data is determined to be the first card usage feature;

[0084] If the target customer is determined to be an active customer who has bound a credit card to an application, the corresponding feature data is determined to be the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature;

[0085] If it is determined that the target customer is a credit card application churn warning customer, the corresponding feature data is determined to be the third card usage feature, the second application behavior feature, and the second card usage and behavior composite feature;

[0086] If it is determined that the type of target customer is a credit card application churn customer, the corresponding feature data is determined to be the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature.

[0087] In a possible implementation, the processing module is specifically configured to:

[0088] Based on the characteristic data, determine whether the target customer is a recommended customer in the credit card application binding process;

[0089] In response to the target customer being a push customer for a credit card application binding phase, an activity push strategy for the credit card application binding phase is determined.

[0090] In one possible implementation, determining whether the target customer is a push customer in the credit card application binding process based on the characteristic data includes:

[0091] Inputting the first card usage feature into a card binding propensity prediction model, training the card binding propensity prediction model until convergence, and using the card binding propensity prediction model to determine a card binding propensity score corresponding to the target customer during the card binding process of the credit card application. The card binding propensity prediction model that has been trained to convergence is stored in a card binding module of the credit card application.

[0092] Based on the card binding propensity score, determine whether the target customer is a recommended customer for the credit card application card binding process.

[0093] In a possible implementation, before inputting the first card usage feature into the card binding tendency prediction model, the processing module is further configured to:

[0094] Based on the churn customer recall module, obtain the churn recall scores corresponding to the credit card application churn customers among the target customers;

[0095] Based on the churn recall score, determine the non-churn recall intention customers and delete them from the target customers.

[0096] In a possible implementation, the processing module is specifically configured to:

[0097] Determine the card binding tendency level based on the card binding tendency score;

[0098] In response to the card binding tendency level being the preset information push level, determining the target customer as a push customer for the credit card application program in the card binding process;

[0099] In response to the card binding tendency level not being the preset information push level, it is determined that the target customer is not a push customer in the card binding link of the credit card application.

[0100] In a possible implementation, the processing module is specifically configured to:

[0101] Determine the corresponding promotion channel based on the target customers' card binding tendency level;

[0102] Obtain customer preference tags based on active customer maintenance in credit card applications;

[0103] Based on the corresponding activity push channels and customer preference tags, determine the activity push strategy for the corresponding card application binding process.

[0104] Push corresponding activity information to target customer terminals according to the corresponding activity push strategy, including:

[0105] Push activity information that matches the target customer's preference tags to the target customer terminal based on the corresponding activity push channel.

[0106] In a possible implementation, the processing module is specifically configured to:

[0107] Determining at least one of a target customer's credit card application active behavior tag and a customer preference tag based on the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature;

[0108] Based on the credit card application binding process, obtain the target customer's activity push channel;

[0109] Determine a corresponding activity push strategy based on at least one of the target customer's credit card application active behavior tag, the customer preference tag, and / or the activity push channel;

[0110] Push corresponding activity information to target customer terminals according to the corresponding activity push strategy, including:

[0111] Pushing activity information to the target customer terminal via the credit card application based on at least one of the target customer's credit card application active behavior tag and customer preference tag;

[0112] And / or, at least one of the target customer's credit card application active behavior tag and the customer preference tag is used to push activity information to the target customer terminal through the target customer's activity push channel.

[0113] In a possible implementation, the processing module is specifically configured to:

[0114] Based on the active customer maintenance phase of the credit card application, obtain the active behavior label of the credit card application;

[0115] Inputting the third card usage feature, the second application behavior feature, the second card usage and behavior composite feature, and the credit card application active behavior label into a churn warning model, training the churn warning model until convergence, and using the churn warning model that has been trained to convergence to determine a churn warning score corresponding to the target customer in the credit card application customer churn warning phase. The churn warning model that has been trained to convergence is located in a customer churn warning module.

[0116] Based on the third card usage feature, obtain the target customer's card churn category label. The target customer's card usage flow type label is used to indicate the probability that the target customer will not use the corresponding credit card;

[0117] Determine the churn customer group to which the target customer belongs based on their churn warning score and card usage churn category label in the credit card application's customer churn warning process;

[0118] Based on the churn customer group to which the target customer belongs, determine the corresponding activity push strategy in the customer churn warning link of the credit card application.

[0119] In a possible implementation, the processing module is specifically configured to:

[0120] If it is determined that the target customer belongs to the first-level churn customer group, the activity push strategy corresponding to the customer churn warning phase of the credit card application is determined to be the first-level churn push strategy;

[0121] If it is determined that the target customer belongs to the second-level churn customer group, the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the second-level churn push strategy;

[0122] If it is determined that the target customer belongs to the third-level churn customer group, then the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the third-level churn push strategy;

[0123] If it is determined that the target customer belongs to the fourth-level churn customer group, then the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the fourth-level churn push strategy;

[0124] In a possible implementation, the processing module is specifically configured to:

[0125] If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to the first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to the second threshold, then the target customer belongs to the first-level churn customer group;

[0126] If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to the first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than the second threshold, then the target customer belongs to the second-level churn customer group;

[0127] If the target customer's churn warning score in the credit card application's customer churn warning phase is less than the first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to the second threshold, then the target customer belongs to the third-level churn customer group;

[0128] If the target customer's churn warning score corresponding to the credit card application's customer churn warning link is less than the first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than the second threshold, then the target customer belongs to the fourth-level churn customer group.

[0129] In a possible implementation, the processing module is specifically configured to:

[0130] inputting the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature into a churn recall model, training the churn recall model until convergence, and using the churn recall model that has been trained to convergence to determine a churn recall score corresponding to the target customer in the credit card application churn customer recall phase, wherein the churn recall model that has been trained to convergence is located in a churn customer recall module;

[0131] Obtain the target customer's churn warning score from the customer churn warning module;

[0132] Determine the recall customer group to which the target customer belongs based on their churn recall score and churn warning score in the credit card application's churn recall phase;

[0133] Obtaining the target customer's card transaction preference tag based on the fourth card usage characteristic;

[0134] Based on at least one of the recall customer group to which the target customer belongs and the card transaction preference tag, a corresponding activity push strategy for recalling churned customers in a credit card application is determined.

[0135] In one possible implementation, determining the recall customer group to which the target customer belongs based on the target customer's churn recall score and churn warning score corresponding to the churn recall phase of the credit card application includes:

[0136] If the target customer's corresponding churn recall score is greater than or equal to the third threshold, and the churn warning score is greater than or equal to the fourth threshold, the target customer's recall customer group is determined to be the first-level recall customer group;

[0137] If the churn recall score corresponding to the target customer is greater than or equal to the third threshold, and the churn warning score is less than the fourth threshold, the target customer is determined to belong to the second-level recall customer group;

[0138] If the churn recall score corresponding to the target customer is less than the third threshold, and the churn recall score corresponding to the target customer is greater than or equal to the fifth threshold, the target customer is determined to belong to the third-level recall customer group;

[0139] If the churn recall score corresponding to the target customer is less than the fifth threshold, the target customer is determined to belong to the fourth-level recall customer group;

[0140] In a possible implementation, the processing module is specifically configured to:

[0141] If it is determined that the churned customer group to which the target customer belongs is a first-level recall customer group, then the activity push strategy corresponding to the churned customer recall phase of the credit card application is determined to be a first-level recall push strategy, which includes pushing activity information to the target customer;

[0142] If it is determined that the target customer belongs to a second-level recall customer group, then the corresponding activity push strategy for the lapsed customer recall phase of the credit card application is determined to be a second-level recall push strategy. The second-level push strategy includes pushing activity information matching the card transaction preference tag to the target customer.

[0143] If it is determined that the target customer belongs to the third-level recall customer group, the corresponding activity push strategy for the churn customer recall phase of the credit card application is determined to be the third-level recall push strategy. The third-level recall push strategy includes pushing activity information matching the card transaction preference tag to the target customer.

[0144] If it is determined that the churned customer group to which the target customer belongs is the fourth-level recall customer group, then the activity push strategy corresponding to the churned customer recall link of the credit card application is determined to be the fourth-level recall push strategy.

[0145] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a memory, a processor;

[0146] Memory stores computer-executable instructions;

[0147] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0148] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementation methods of the first aspect.

[0149] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect.

[0150] Embodiments of the present application provide a data processing method, electronic device, storage medium, and program product. These methods determine the type of target customer to be processed. Based on this type, they then identify characteristic data associated with this type of customer. Based on this characteristic data, they determine the target customer's current phase, and then push corresponding activity information to the target customer according to the corresponding activity push strategy. Because the target customer is at least one of a credit card application unbound customer, an active credit card application bound customer, a credit card application churn warning customer, or a credit card application churn customer, and the corresponding activity push phase is one of the credit card application binding phase, the credit card application active customer maintenance phase, the credit card application churn warning phase, or the credit card application churn customer recall phase, the push service covers the entire lifecycle of a customer's credit card and credit card application use. Regardless of the target customer's type or lifecycle phase, they can receive appropriate activity push services. Furthermore, when determining the target customer's characteristic data, not only credit card application behavior characteristics but also credit card usage characteristics, as well as combined card usage and behavior characteristics, are considered. This makes the determined characteristics more comprehensive and integrated, enabling more accurate push of activity information to the target customer. BRIEF DESCRIPTION OF THE DRAWINGS

[0151] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0152] Figure 1 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 1 ;

[0153] Figure 2 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 2 ;

[0154] Figure 3 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 3 ;

[0155] Figure 4 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 4 ;

[0156] Figure 5 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 5 ;

[0157] Figure 6 A schematic diagram of data transmission between different links provided in an embodiment of the present application;

[0158] Figure 7 A schematic diagram of the structure of a data processing device provided in an embodiment of the present application;

[0159] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0160] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0161] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0162] In the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of information such as financial data or user data involved comply with the provisions of relevant laws and regulations, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0163] The credit card customer lifecycle refers to the entire process from when a customer applies for a credit card, through activation, use, and ultimately cancellation or renewal. The credit card customer lifecycle is defined by four key stages: activation, engagement, churn, and reactivation.

[0164] The application (APP) customer lifecycle refers to the process from the time a customer first encounters a credit card app to the time they eventually stop using the app. The app customer lifecycle is defined as the following four key stages: card registration and binding, active operations, churn warning, and churn recall.

[0165] In the era of mobile internet, apps are not only crucial channels for credit card institutions to connect businesses with customers, but also crucial platforms for cultivating customer relationships and enhancing customer loyalty. Managing app customers plays a crucial role in enhancing engagement and trust between customers and credit card institutions. The importance of app customer lifecycle management is self-evident. Through a deep understanding and effective management of the app customer lifecycle, credit card institutions can more accurately grasp customer needs at every stage, from engagement, attraction, conversion, retention, and value-added, continuously improving the customer experience and maximizing customer value. This not only helps to improve customer loyalty and retention, but also provides companies with sustained growth momentum, ensuring they maintain their leading position in the fiercely competitive market.

[0166] In some embodiments, an activity push strategy is formulated based on the customer's behavioral trajectory and customer profile in the APP itself as a data source, thereby pushing activity information to the customer; however, this embodiment fails to fully mine the behavioral data of the customer's other financial attributes, making it difficult to formulate a comprehensive activity push strategy, and thus unable to accurately push activity information to the customer.

[0167] In some embodiments, a corresponding activity push strategy is formulated and implemented only for a single stage of the APP customer life cycle, thereby pushing activity information to customers, which fails to fully cover the APP customer life cycle.

[0168] Therefore, to address the difficulty in accurately pushing campaign information to customers, in addition to identifying credit card application behavioral characteristics, one can also determine credit card usage characteristics, as well as combined card usage and behavioral characteristics. Based on these characteristics, corresponding campaign strategies can be developed to more accurately push campaign information to target customers. Furthermore, to address the technical issue of incomplete coverage of the app customer lifecycle, customer types can be pre-configured as either unlinked credit card application customers, active credit card application linked credit card application customers, churn warning credit card application customers, or churn credit card application customers. Furthermore, campaign push phases can be configured as either credit card application linked credit card application, active credit card application maintenance phase, churn warning credit card application, or churn recall phase. Any customer can then be categorized into any of these types, and a corresponding campaign push strategy can be determined for the push phase corresponding to that type, thereby completing the push of campaign information for that phase. This allows push services to cover the entire lifecycle of a customer's credit card and credit card application usage, ensuring that target customers, regardless of their type or lifecycle phase, receive the appropriate campaign push service.

[0169] Therefore, the present application proposes a data processing method, electronic device, storage medium, and program product. The method determines the type of target customer to be processed. Based on this type, characteristic data related to this type of customer is found. Based on this characteristic data, the target customer's current stage is determined, and corresponding activity information is pushed to the target customer according to the activity push strategy for the corresponding stage. Since the target customer is at least one of a credit card application unbound customer, a credit card application bound active customer, a credit card application churn warning customer, or a credit card application churn customer, and the corresponding activity push stage is one of the credit card application binding stage, the credit card application active customer maintenance stage, the credit card application churn warning stage, and the credit card application churn customer recall stage, the push service covers the entire lifecycle of a customer's credit card and credit card application use. Regardless of the target customer's type or stage in the lifecycle, they can receive appropriate activity push services. Furthermore, when determining the target customer's characteristic data, not only credit card application behavior characteristics but also credit card usage characteristics, as well as combined card usage and behavior characteristics, are considered. This makes the determined characteristics more comprehensive and integrated, enabling more accurate push of activity information to the target customer.

[0170] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0171] Figure 1 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 1 .like Figure 1 As shown, the method includes the following steps:

[0172] S101. Determine the type of target customers to be processed.

[0173] The target customer type is at least one of the following: a customer who has not bound a card to a credit card application, an active customer who has bound a card to a credit card application, a customer who has been warned of churn by a credit card application, or a customer who has churned by a credit card application.

[0174] In some embodiments, the type of the target customer is determined based on at least one of whether the target customer has bound a corresponding credit card in a credit card application and whether the target customer has launched the credit card application within a preset time period, including at least one of the following situations:

[0175] Case 1: If the target customer has not bound a corresponding credit card in the credit card application, it is determined that the type of the target customer is a customer who has not bound a card to the credit card application.

[0176] Case 2: If the target customer binds the corresponding credit card in the credit card application and the target customer launches the credit card application within a preset time period, the type of the target customer is determined to be an active customer bound to the credit card application.

[0177] Case 3: If the target customer binds the corresponding credit card in the credit card application and the target customer activates the credit card application within a preset time period, the type of the target customer is determined to be a credit card application churn warning customer.

[0178] In a possible implementation, based on actual needs of the credit card business, the preset time period in case 2 and the preset time period in case 3 may be the same preset time period, which is not limited in this embodiment.

[0179] For example, if the preset time period of situation 2 and the preset time period of situation 3 are both within the past month, then the credit card application active customer maintenance link and the credit card application customer churn warning link can be simultaneously carried out for customers who have activated the credit card application within the past month, thereby improving the activity and retention rate at the same time; if the preset time period of situation 2 is within the past year and the preset time period of situation 3 is within the past month, then a larger range of credit card application active customer maintenance can be carried out for customers who have activated the credit card application within the past year, and a more timely customer churn warning can be issued for customers who have activated the credit card application within the past month, thereby improving the retention rate.

[0180] Case 4: If the target customer does not launch the credit card application within the preset time period, the target customer is determined to be a credit card application churn customer.

[0181] In a possible implementation manner, the preset time period in case 4 and the preset time period in case 3 are the same preset time period.

[0182] In some embodiments, the credit card application lost customer may be a target customer whose corresponding credit card is bound in the credit card application, or may be a target customer whose corresponding credit card is not bound in the credit card application.

[0183] S102: Determine corresponding feature data according to the type of target customer.

[0184] The characteristic data includes at least one of the following: card usage characteristics, credit card application behavior characteristics, and a combination of card usage and behavior characteristics. The combination of card usage and behavior characteristics is determined based on the card usage characteristics and credit card application behavior characteristics. In some embodiments, the corresponding characteristic data is determined based on the type of target customer, including at least one of the following situations:

[0185] Case 1: If it is determined that the target customer is a customer who has not bound a card to a credit card application, the corresponding feature data is determined to be the first card usage feature.

[0186] Illustratively, the first card usage feature includes but is not limited to customer billing information, transaction information, key card usage nodes, etc.

[0187] Case 2: If the target customer type is determined to be an active customer who has bound the card to a credit card application, the corresponding feature data is determined to be the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature.

[0188] Case 3: If the target customer type is determined to be a credit card application churn warning customer, the corresponding feature data is determined to be the third card usage feature, the second application behavior feature, and the second card usage and behavior composite feature.

[0189] For example, the second application behavior feature includes but is not limited to the number of days the APP is logged in, the click rate of the APP page, etc.; the second card usage and behavior composite feature includes but is not limited to the intersection of whether the card is used by an active customer × the number of APP logins, whether the card is used by a churned customer × the number of visits to the APP security settings page, etc.

[0190] Case 4: If the target customer type is determined to be a credit card application churn customer, the corresponding feature data is determined to be the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature.

[0191] For example, the fourth card usage feature includes but is not limited to the card repayment rate; the third application behavior feature includes but is not limited to the number of days the APP has not been logged in, the APP login method, etc.; the third card usage and behavior composite feature includes but is not limited to the intersection of whether the card is lost and the number of visits to the APP financial page, etc.

[0192] Specifically, by setting key tracking points in the application, the behavioral characteristics of the credit card application are obtained, such as startup login, function usage, transaction payment, behavioral trajectory, etc.; based on customer credit card usage data, such as billing data, transaction data, etc., card usage characteristics are obtained.

[0193] For example, card usage characteristics include but are not limited to customer credit card bill characteristics, customer credit card transaction characteristics, key card usage node characteristics, etc.; credit card application behavior characteristics include but are not limited to startup login, behavior trajectory, function usage, transaction payment, device system, etc.

[0194] In a possible embodiment, the composite card usage and behavior feature is determined based on the card usage feature and the credit card application behavior feature through the following steps S11-S13, including:

[0195] Step S11: obtaining first basic features by cleaning and converting card usage features and credit card application behavior features.

[0196] Among them, cleaning and conversion include operations such as missing value processing, outlier processing, data type conversion, time series conversion, etc.

[0197] For example, the first basic feature may be features such as APP login duration, APP page click rate, APP activity participation, credit card consumption amount, and credit card balance.

[0198] Step S12: Based on the first basic feature that has been cleaned and converted, the first basic feature is processed through variable derivation to obtain a first intermediate feature.

[0199] For example, variable derivation can be performed by measuring cyclical scale, cyclical trend changes, non-cyclical states, etc. The first intermediate features include intermediate card usage features and intermediate credit card application behavior features. For example, the first intermediate features can include 90-day active credit card users, card spending amount, APP customer churn, APP login times, etc.

[0200] Among them, the method of measuring cyclical scale includes calculating the total value, average value, ratio, extreme value, etc. for the frequency of behaviors and the amount involved within a time window; the method of measuring cyclical trend changes includes the rolling change trend of indicators in equal-time windows, and calculating the continuous growth, continuous decline, and longest duration for the frequency of behaviors and the amount involved within a time window; the method of measuring non-cyclical status includes calculating the first occurrence time, the last occurrence time, and the most frequent behavior of a certain behavior.

[0201] Step S13: Determine the composite features of card usage and behavior by combining the intermediate features of card usage and the intermediate features of credit card application behavior in pairs.

[0202] Among them, the composite features are composed of dimensions such as APP access depth and breadth, APP high-frequency access scenarios, credit card transaction scale, and customer credit card life cycle status, which completely cover the application customer life cycle and credit card customer life cycle.

[0203] Among them, the intermediate features of card usage can only be combined with the intermediate features of credit card application behavior in pairs, and the intermediate features of credit card application behavior can only be combined with the intermediate features of card usage in pairs.

[0204] Exemplarily, the composite features of card usage and behavior can be expressed as intermediate features of card usage × intermediate features of credit card application behavior (or, intermediate features of credit card application behavior × intermediate features of card usage), including but not limited to the intersection of whether the customer is an active card user × the number of APP logins, whether the customer has churned the card × the number of visits to the APP security settings page, whether the customer has a history of card binding on the APP × the number of active card usage days, and whether the customer has churned the card × the number of visits to the APP financial page.

[0205] S103: Determine the activity push strategy for the corresponding link based on the feature data.

[0206] Among them, the corresponding links are one of the following: credit card application binding link, credit card application active customer maintenance link, credit card application customer churn warning link, and credit card application churn customer recall link.

[0207] Specifically, based on the four stages of the application customer life cycle, including registration and card binding, active operation, churn warning, and churn recall, four corresponding links are established, and the activity push strategy of the corresponding link is determined based on the characteristic data to fully cover the application customer life cycle.

[0208] S104: Push corresponding activity information to target client terminals according to corresponding activity push strategies.

[0209] In a possible implementation, the corresponding activity push strategy and corresponding activity information are pushed to the account manager terminal so that the account manager can reach customers based on actual experience, enhance customer relationships and improve customer satisfaction.

[0210] The present embodiment provides a data processing method. This method determines the type of target customer to be processed. Based on this type, characteristic data related to this type of customer is then found. Based on this characteristic data, the target customer's current stage is determined, and corresponding activity information is then pushed to the target customer according to the activity push strategy for the corresponding stage. Since the target customer is at least one of a credit card application unbound customer, a credit card application bound active customer, a credit card application churn warning customer, or a credit card application churn customer, and the corresponding activity push stage is one of the credit card application binding stage, the credit card application active customer maintenance stage, the credit card application churn warning stage, or the credit card application churn customer recall stage, the push service covers the entire lifecycle of a customer's credit card and credit card application use. Regardless of the target customer's type or stage in the lifecycle, they can receive appropriate activity push services. Furthermore, when determining the target customer's characteristic data, not only credit card application behavior characteristics but also credit card usage characteristics, as well as combined card usage and behavior characteristics, are considered. This makes the determined characteristics more comprehensive and integrated, enabling more accurate push of activity information to the target customer.

[0211] Figure 2 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 2 .like Figure 2 As shown, the method includes the following steps:

[0212] S201: Determine that the type of the target customer to be processed is a customer whose credit card application has not been bound to a card.

[0213] S202: Determine corresponding feature data according to the type of target customer.

[0214] It should be noted that, in the embodiment of the present application, the specific implementation process of step S201 and step S202 can refer to the specific implementation process of the above-mentioned step S101 and step S102, which will not be repeated here.

[0215] S203: Determine, based on the characteristic data, whether the target customer is a push customer in the credit card application binding process.

[0216] In some embodiments, the first card usage feature is input into a card binding tendency prediction model, the card binding tendency prediction model is trained until convergence, and the card binding tendency prediction model that has been trained to convergence is used to determine the card binding tendency score corresponding to the target customer in the credit card application binding link. The card binding tendency prediction model that has been trained to convergence is located in a credit card application binding module, and the credit card application binding module is used to execute the credit card application binding link.

[0217] In one possible implementation, the first card usage feature is subjected to variable processing, such as missing value processing, outlier processing, data type conversion, time series conversion, etc., and the first card usage feature after variable processing is input into a card binding tendency prediction model that has been trained to convergence, and the card binding tendency prediction model is used to determine the target customer's corresponding card binding tendency score in the card binding stage of the credit card application.

[0218] In one possible implementation, a training data set is obtained, the training data set is input into a card binding tendency prediction model, a high-order machine learning algorithm of random forest is applied to train the card binding tendency prediction model until convergence, and a card binding tendency prediction model that has been trained to convergence is obtained.

[0219] In some optional embodiments, before inputting the first card usage feature into the card binding tendency prediction model, the following steps S21-S22 may be further included:

[0220] Step S21: Based on the churned customer recall module, obtain churn recall scores corresponding to churned credit card application customers among target customers.

[0221] Specifically, if the target customer has not bound the corresponding credit card in the credit card application, and the target customer has not started the credit card application within the preset time period, then the type of the target customer is a credit card application unbound customer and a credit card application churned customer. Therefore, based on the churned customer recall module, the churn recall score corresponding to the credit card application churned customers among the target customer can be obtained.

[0222] Step S22: Based on the churn recall score, determine the non-churn recall-willing customers, and delete the non-churn recall-willing customers from the target customers.

[0223] Specifically, if the target customer's churn recall score is less than the third threshold, the target customer is determined to be a non-churn recall-willing customer and is deleted from the target customers to optimize data analysis and card binding tendency prediction model performance.

[0224] Based on the card binding propensity score, determine whether the target customer is a recommended customer for the credit card application card binding process.

[0225] Specifically, based on the card binding tendency score, the card binding tendency level is determined; in response to the card binding tendency level being the preset information push level, the target customer is determined to be a push customer in the card binding link of the credit card application; in response to the card binding tendency level not being the preset information push level, the target customer is determined not to be a push customer in the card binding link of the credit card application.

[0226] Among them, the card binding propensity score is used to indicate the probability of the target customer binding the card in the credit card application, and the preset information push level is used to indicate that the target customer is a push customer who needs to be pushed activity information to encourage the target customer to bind the card.

[0227] For example, the card binding propensity score may be a value between 0 and 1, where a larger value indicates a higher card binding propensity of the target customer. Based on the card binding response score, the card binding propensity level may be determined, which may include the following five situations:

[0228] Case 1: If the card binding response score is greater than or equal to 0.8, it indicates that the target customer has a high tendency to bind a card. The target customer can bind a card on their own and there is no need to push activity information to encourage the target customer to bind a card. In this case, the target customer's card binding tendency level is determined to be high and is not the preset information push level. Therefore, it is determined that the target customer is not a push customer for the card binding process of the credit card application.

[0229] Case 2: If the card binding response score is less than 0.8 and greater than or equal to 0.6, the target customer's card binding tendency level is determined to be medium-high and is at the preset information push level, and the target customer is further determined to be a push customer in the card binding stage of the credit card application.

[0230] Case 3: If the card binding response score is less than 0.6 and greater than or equal to 0.4, the target customer's card binding tendency level is determined to be medium, and is at the preset information push level, and the target customer is further determined to be a push customer in the card binding stage of the credit card application.

[0231] Case 4: If the card binding response score is less than 0.4 and greater than or equal to 0.2, the target customer's card binding tendency level is determined to be medium-low, and is at the preset information push level, and the target customer is further determined to be a push customer in the card binding stage of the credit card application.

[0232] Case 5: If the card binding response score is less than 0.2, it means that the target customer has a low binding tendency. Even if the activity information is pushed, it is difficult to encourage the target customer to bind the card. In this case, the target customer's card binding tendency level is determined to be a low card binding tendency level, and it is not the preset information push level. Therefore, it is determined that the target customer is not a push customer in the card binding link of the credit card application.

[0233] In one possible implementation, after determining whether a target customer is a recommended customer for the credit card application binding process based on the card binding propensity score, a churn recall module can be used to obtain the corresponding churn recall score for any churned credit card application customers among the recommended customers. Based on the churn recall score, non-churn recall-intended customers are identified and removed from the recommended customers list.

[0234] S204: In response to the target customer being a push customer for the credit card application binding phase, determine an activity push strategy for the credit card application binding phase.

[0235] Specifically, based on the target customers' card binding tendency level, the corresponding activity push channels are determined. The activity push channels include but are not limited to SMS, AI phone calls and other channels to reach the target customers; based on the active customer maintenance link of the credit card application, customer preference tags are obtained. Customer preference tags include but are not limited to financial management, digital, etc.; based on the corresponding activity push channels and customer preference tags, the activity push strategy for the corresponding credit card application card binding link is determined.

[0236] For example, if the target customer's card binding tendency level is medium and the customer preference label is digital, it means that the target customer's card binding tendency is weak, and an activity push channel with stronger reach intensity can be adopted. Then, the activity push strategy for the corresponding credit card application binding link includes the activity channel being an artificial intelligence phone, and the activity information in the phone is designed centered on the customer preference label, namely digital.

[0237] S205: Push corresponding activity information to target client terminals according to corresponding activity push strategies.

[0238] Specifically, the activity information matching the target customer's preference tag is pushed to the target customer terminal based on the corresponding activity push channel.

[0239] The embodiment of the present application proposes a data processing method. It is determined that the type of target customer to be processed is a customer who has not bound a card to a credit card application. According to the type of target customer, the corresponding characteristic data is determined. According to the characteristic data, it is determined whether the target customer is a push customer in the card binding link of the credit card application. In response to the target customer being a push customer in the card binding link of the credit card application, the activity push strategy for the card binding link of the credit card application is determined, and the corresponding activity information is pushed to the target customer terminal according to the corresponding activity push strategy. In case the target customer is a customer who has bound a card to a credit card application, since the target customer has not bound a card to the credit card application, the credit card usage data of the target customer can be used as a data source to analyze the characteristics of the customer. Then, according to the characteristic data of the target customer, activity push strategies with different motivation levels can be designed to encourage target customers with different characteristics to bind a card. Activity information can also be designed in combination with customer preference tags to attract the interest of target customers.

[0240] Figure 3 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 3 .like Figure 3 As shown, the method includes the following steps:

[0241] S301: Determine that the type of the target customer to be processed is an active customer who has bound a credit card application.

[0242] S302: Determine corresponding feature data according to the type of target customer.

[0243] It should be noted that, in the embodiment of the present application, the specific implementation process of step S301 and step S302 can refer to the specific implementation process of the above-mentioned step S101 and step S102, which will not be described in detail here.

[0244] S303. Determine at least one of the target customer's credit card application active behavior tag and the customer preference tag based on the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature.

[0245] In one possible implementation, the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature are input into a credit card application active behavior model, and based on the credit card application active behavior model, the target customer's credit card application active behavior label is determined.

[0246] In a possible implementation, the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature are input into a customer preference model, and based on the customer preference model, a customer preference tag of the target customer is determined.

[0247] S304: Based on the credit card application binding process, obtain the target customer's activity push channel.

[0248] Specifically, if the target customer binds the card through an activity push channel during the card binding process of the credit card application, the activity push channel of the target customer is obtained.

[0249] It should be noted that since some target customers have not bound their cards through any activity push channel, step S304 is an optional implementation step.

[0250] S305: Determine a corresponding activity push strategy based on at least one of the target customer's credit card application active behavior tag, the customer preference tag, and / or the activity push channel.

[0251] For example, if the target customer's credit card application active behavior tag includes food, the customer preference tag includes digital, and the activity push channel is SMS, then the corresponding activity push strategy is determined to push activity information via SMS, and the activity information is designed centered on digital products.

[0252] S306: Push corresponding activity information to the target client terminal according to the corresponding activity push strategy.

[0253] Specifically, based on at least one of the target customer's credit card application active behavior tag and customer preference tag, activity information is pushed to the target customer terminal through the credit card application; and / or, based on at least one of the target customer's credit card application active behavior tag and customer preference tag, activity information is pushed to the target customer terminal through the target customer's activity push channel.

[0254] For example, within a credit card application, if the target customer's credit card application active behavior tags include food, and the customer preference tags include digital products, then information related to digital products will be pushed to the target customer terminal through the food page in the credit card application to increase the customer's click-through rate; outside the credit card application, if the target customer's credit card application active behavior tags include food, and the customer preference tags include digital products, and the target customer's activity push channel is SMS, then food- and digital-related activity information will be sent to the target customer terminal via SMS to achieve differentiated channel customer maintenance and enhance the credit card application's exposure and attractiveness to customers.

[0255] The present application embodiment provides a data processing method. The method includes determining that the type of target customer to be processed is an active customer bound to a credit card application, determining corresponding feature data based on the type of the target customer, determining at least one of the target customer's credit card application active behavior tag and customer preference tag based on a second card usage feature, a first application behavior feature, and a first card usage and behavior composite feature, obtaining an activity push channel for the target customer based on the credit card application binding process, determining a corresponding activity push strategy based on at least one of the target customer's credit card application active behavior tag and customer preference tag, and / or the activity push channel, and pushing corresponding activity information to the target customer terminal in accordance with the corresponding activity push strategy. The target customer type is an active customer who has bound his card to a credit card application. Since the target customer uses both a credit card application (and binds the card) and a credit card, the credit card usage data and the credit card application behavior data can be used as data sources, thereby comprehensively analyzing the customer's card usage characteristics, credit card application behavior characteristics, and card usage and behavior composite characteristics, which can make the determined characteristics more comprehensive and integrated, and thus can more accurately push activity information to the target customer; and by combining at least one of the target customer's credit card application active behavior tag and customer preference tag, the corresponding activity push strategy can be determined, which can accurately attract the customer's interest and increase the customer's activity.

[0256] Figure 4 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 4 .like Figure 4 As shown, the method includes the following steps:

[0257] S401: Determine that the type of the target customer to be processed is a credit card application churn warning customer.

[0258] S402: Determine corresponding feature data according to the type of target customer.

[0259] It should be noted that, in the embodiment of the present application, the specific implementation process of step S401 and step S402 can refer to the specific implementation process of the above-mentioned step S101 and step S102, which will not be repeated here.

[0260] S403: Based on the active customer maintenance phase of the credit card application, obtain the active behavior tag of the credit card application.

[0261] S404. Input the third card usage feature, the second application behavior feature, the second card usage and behavior composite feature, and the credit card application active behavior label into the churn warning model, train the churn warning model until convergence, and use the churn warning model that has been trained to convergence to determine the target customer's corresponding churn warning score in the credit card application customer churn warning link.

[0262] Among them, the churn warning model that has been trained to convergence is located in the customer churn warning module. The customer churn warning module is used to execute the customer churn warning link of the credit card application. The churn warning score is used to indicate the probability of credit card application customer churn within a preset time period, and the churn warning score will be input into the credit card application churn customer recall link to support the churn recall link in determining the activity push strategy.

[0263] In one possible implementation, a training data set is obtained, the training data set is input into a card binding tendency prediction model, a gradient boosting algorithm is applied to train the card binding tendency prediction model until convergence, and a card binding tendency prediction model that has been trained to convergence is obtained.

[0264] S405. Obtain a card usage churn category label of the target customer based on the third card usage feature.

[0265] The target customer's card churn type label is used to indicate the probability that the target customer does not use the corresponding credit card.

[0266] S406: Determine the churn customer group to which the target customer belongs based on the churn warning score corresponding to the target customer in the credit card application's customer churn warning link and the card usage churn category label.

[0267] Specifically, to focus resources on high-value customers who are more likely to churn, target customers need to be divided into different customer groups. Therefore, based on the target customer's churn warning score and card churn category label in the credit card application's customer churn warning phase, the churn customer group to which the target customer belongs is determined, including the following four situations:

[0268] Case 1: If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to the first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to the second threshold, then the target customer belongs to the first-level churn customer group, which is a customer group with a high churn tendency.

[0269] Case 2: If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to the first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than the second threshold, then the target customer belongs to the second-level churn customer group, which is a customer group with a higher churn tendency.

[0270] Case 3: If the target customer's churn warning score in the credit card application's customer churn warning section is less than the first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to the second threshold, then the target customer belongs to the third-level churn customer group, which is a customer group with a lower churn tendency.

[0271] Case 4: If the target customer's churn warning score in the credit card application's customer churn warning section is less than the first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than the second threshold, then the target customer belongs to the fourth-level churn customer group, and the third-level churn customer group is the customer group with a low churn tendency.

[0272] S407: Based on the churn customer group to which the target customer belongs, determine an activity push strategy corresponding to the customer churn warning phase of the credit card application.

[0273] There are four situations:

[0274] Case 1: If it is determined that the target customer belongs to a first-level churn customer group, then the activity push strategy corresponding to the customer churn warning link of the credit card application is determined to be a first-level churn push strategy.

[0275] Case 2: If it is determined that the target customer belongs to a second-level churn customer group, then the activity push strategy corresponding to the customer churn warning link of the credit card application is determined to be the second-level churn push strategy.

[0276] Case 3: If it is determined that the target customer belongs to the third-level churn customer group, then the activity push strategy corresponding to the customer churn warning link of the credit card application is determined to be the third-level churn push strategy.

[0277] Case 4: If it is determined that the target customer belongs to the fourth-level churn customer group, then the activity push strategy corresponding to the customer churn warning link of the credit card application is determined to be the fourth-level churn push strategy.

[0278] Among them, the first-level churn push strategy can be a proactive and high-cost strategy, such as artificial intelligence phone calls; the second-level churn push strategy can be a proactive and medium-cost strategy; the third-level churn push strategy can be a high-frequency and low-cost strategy, such as text messages; the fourth-level churn push strategy can be a low-frequency and low-cost strategy, such as pushing activities within credit card applications.

[0279] S408: Push corresponding activity information to target client terminals according to corresponding activity push strategies.

[0280] The present application embodiment provides a data processing method. The method includes determining that the type of target customer to be processed is a credit card application churn warning customer, determining corresponding feature data based on the type of target customer, obtaining a credit card application active behavior label based on the credit card application active customer maintenance phase, inputting a third card usage feature, a second application behavior feature, a second card usage and behavior composite feature, and the credit card application active behavior label into a churn warning model, training the churn warning model until convergence, and using the churn warning model trained to convergence to determine a churn warning score corresponding to the target customer in the credit card application customer churn warning phase. Based on the third card usage feature, obtaining a card usage churn category label for the target customer, determining the churn customer group to which the target customer belongs based on the churn warning score and card usage churn category label corresponding to the credit card application customer churn warning phase, determining an activity push strategy corresponding to the credit card application customer churn warning phase based on the churn customer group to which the target customer belongs, and pushing corresponding activity information to the target customer terminal according to the corresponding activity push strategy. The target customer type is a credit card application churn warning customer. Since the target customer uses both a credit card application and a credit card, the credit card usage data and the credit card application behavior data can be used as data sources. This allows for a comprehensive analysis of the customer's card usage characteristics, credit card application behavior characteristics, and combined card usage and behavior characteristics. This allows the identified characteristics to more comprehensively and comprehensively demonstrate the target customer's churn tendency. Different activity push strategies can be formulated for target customers with different churn tendency levels, and activity information can be accurately pushed to intervene in the customer's churn tendency.

[0281] Figure 5 A schematic diagram of a data processing method provided in an embodiment of the present application Figure 5 .like Figure 5 As shown, the method includes the following steps:

[0282] S501: Determine that the type of the target customer to be processed is a credit card application churn customer.

[0283] S502: Determine corresponding feature data according to the type of target customer.

[0284] It should be noted that, in the embodiment of the present application, the specific implementation process of step S501 and step S502 can refer to the specific implementation process of step S101 and step S102 mentioned above, and will not be described in detail here.

[0285] S503. Input the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature into a churn recall model, train the churn recall model until convergence, and use the churn recall model that has been trained to convergence to determine the churn recall score corresponding to the target customer in the credit card application churn customer recall link.

[0286] In one possible implementation, a training data set is obtained, and the training data set is screened using a random forest machine learning model, information value contribution assessment, and other methods to pre-screen data related to the recall of churned customers in a credit card application program. The screened training data set is input into a churn recall prediction model, and a gradient boosting algorithm is applied to train the churn recall prediction model until convergence, thereby obtaining a churn recall prediction model that has been trained to convergence.

[0287] Among them, the churn recall model that has been trained to convergence is located in the churn customer recall module, which is used to execute the churn customer recall link of the credit card application. The churn recall score is used to indicate the probability of a customer logging into the credit card application within a preset time period, and the churn recall score can be used as input to the credit card application binding link to support the credit card application binding link in determining the activity push strategy.

[0288] S504: Obtain the target customer's churn warning score from the customer churn warning module.

[0289] S505 : Determine the recall customer group to which the target customer belongs based on the churn recall score and churn warning score corresponding to the target customer in the churn customer recall phase of the credit card application program.

[0290] Specifically, there are four situations:

[0291] Case 1: If the churn recall score corresponding to the target customer is greater than or equal to the third threshold, and the churn warning score is greater than or equal to the fourth threshold, then the recall customer group to which the target customer belongs is determined to be the first-level recall customer group, and the first-level recall customer group is a high-recall tendency customer group.

[0292] Case 2: If the churn recall score corresponding to the target customer is greater than or equal to the third threshold, and the churn warning score is less than the fourth threshold, then the recall customer group to which the target customer belongs is determined to be the second-level recall customer group, and the second-level recall customer group is a customer group with a higher recall tendency.

[0293] Case 3: If the churn recall score corresponding to the target customer is less than the third threshold, and the churn recall score corresponding to the target customer is greater than or equal to the fifth threshold, then the recall customer group to which the target customer belongs is determined to be the third-level recall customer group, and the third-level recall customer group is a customer group with a lower recall tendency.

[0294] Case 4: If the churn recall score corresponding to the target customer is less than the fifth threshold, the recall customer group to which the target customer belongs is determined to be the fourth-level recall customer group, and the fourth-level recall customer group is a low-recall tendency customer group.

[0295] S506. Obtain the target customer's card transaction preference tag based on the fourth card usage feature.

[0296] S507: Determine an activity push strategy corresponding to the churned customer recall phase of the credit card application based on at least one of the recall customer group to which the target customer belongs and the card transaction preference tag.

[0297] Specifically, there are four situations:

[0298] Case 1: If it is determined that the lost customer group to which the target customer belongs is the first-level recall customer group, then the activity push strategy corresponding to the lost customer recall link in the credit card application is determined to be the first-level recall push strategy, which includes pushing activity information to the target customer.

[0299] Possibly, the first-level push strategy is more costly and has a strong motivational force.

[0300] Case 2: If it is determined that the target customer belongs to a lost customer group that is a second-level recall customer group, then the activity push strategy corresponding to the lost customer recall link in the credit card application is determined to be a second-level recall push strategy. The second-level push strategy includes pushing activity information that matches the card transaction preference tag to the target customer.

[0301] Possibly, the second level push strategy is moderately costly and moderately motivating.

[0302] For example, if the target customer's card transaction preference label is food, the second-level recall push strategy can be a restaurant discount activity.

[0303] Case 3: If it is determined that the target customer belongs to a lost customer group that is a third-level recall customer group, then the activity push strategy corresponding to the lost customer recall link in the credit card application is determined to be a third-level recall push strategy. The third-level recall push strategy includes pushing activity information that matches the card transaction preference tag to the target customer.

[0304] Possibly, the third level push strategy is less costly and less motivating.

[0305] For example, if the target customer's card transaction preference label is digital, the second-level recall push strategy can be a digital product promotion activity.

[0306] Case 4: If it is determined that the churned customer group to which the target customer belongs is the fourth-level recall customer group, then the activity push strategy corresponding to the churned customer recall link in the credit card application is determined to be the fourth-level recall push strategy.

[0307] S508: Push corresponding activity information to the target client terminal according to the corresponding activity push strategy.

[0308] In some embodiments, since the fourth-level recall customer group has an extremely low recall tendency and a low input-output ratio, no activity push strategy may be applied.

[0309] The present application provides a data processing method. The method includes determining that the target customer to be processed is a credit card application churn customer, determining corresponding feature data based on the target customer type, inputting a fourth card usage feature, a third application behavior feature, and a third card usage and behavior composite feature into a churn recall model, training the churn recall model until convergence, and using the churn recall model trained to convergence to determine a churn recall score corresponding to the target customer in the credit card application churn recall phase. The method also includes obtaining the target customer's churn warning score from a customer churn warning module, determining the target customer's recall customer group based on the target customer's churn recall score and the churn warning score in the credit card application churn recall phase, obtaining a card transaction preference tag for the target customer based on the fourth card usage feature, determining an activity push strategy corresponding to the credit card application churn recall phase based on at least one of the target customer's recall customer group and the card transaction preference tag, and pushing corresponding activity information to the target customer's terminal in accordance with the corresponding activity push strategy. The target customer type is the credit card application churn customer. Since the target customer uses both the credit card application and the credit card, the credit card usage data and the credit card application behavior data can be used as data sources, and then the customer's card usage characteristics, credit card application behavior characteristics, and card usage and behavior composite characteristics can be comprehensively analyzed. The corresponding churn recall score can be determined through the churn recall model, and the churn warning score of the credit card application customer churn warning link can be combined to achieve a more refined division of the target customers, and then determine the activity push strategy with different motivational strengths, which can more accurately push activity information to the target customers to recover churned customers.

[0310] In some embodiments, for different customer groups, corresponding activity push strategies are formulated and implemented according to the corresponding links of different customer groups, so as to push activity information to customers; however, there is a lack of coherent data transmission and mutual integration between the various links, which reduces the efficiency of formulating and implementing activity push strategies.

[0311] Therefore, based on the above embodiments and Figure 6 , Figure 6 The data transmission diagram between different links provided in the embodiment of the present application further illustrates the coherent data transmission and mutual combination between the different links in the embodiment of the present application, including:

[0312] The credit card application binding process obtains the churn recall score transmitted from the churn customer recall process of the credit card application, which is used to screen out customers who are not willing to be recalled, reducing the amount of data that needs to be processed, thereby improving the efficiency of formulating and implementing activity push strategies; and obtains the customer preference tags transmitted from the active customer maintenance process of the credit card application, so that in the process of implementing the activity push strategy, different activity information can be designed in combination with customer preference tags with different preferences, so as to advance the customer's focus and achieve the purpose of attracting customers.

[0313] The active customer maintenance phase of the credit card application obtains the activity recommendation channels delivered from the card binding phase of the credit card application, so as to achieve differentiated channel customer maintenance during the implementation of the activity push strategy and enhance the exposure and attractiveness of the credit card application to customers.

[0314] The credit card application customer churn warning phase obtains the credit card application active behavior tags transmitted from the credit card application active customer maintenance phase, and uses the churn warning model to accurately identify the target customers' churn tendency.

[0315] The credit card application's churn customer recall phase obtains the churn warning score delivered from the credit card application's customer churn warning phase, so as to accurately classify target customers, achieve precise and targeted push of activity information, and thus implement differentiated intervention measures.

[0316] Figure 7 This is a structural diagram of a data processing device provided in an embodiment of the present application. Figure 7 As shown, the data processing device 700 provided in this embodiment includes:

[0317] Determination module 701, for determining the type of target customer to be processed, where the type of target customer is at least one of the following: a customer who has not bound a credit card application, a customer who has an active credit card application, a customer who has been warned of churn in a credit card application, or a customer who has churned in a credit card application;

[0318] The determination module 701 is further configured to determine corresponding characteristic data based on the type of target customer, the characteristic data including at least one of the following: card usage characteristics, credit card application behavior characteristics, and card usage and behavior composite characteristics, wherein the card usage and behavior composite characteristics are determined based on the card usage characteristics and the credit card application behavior characteristics;

[0319] Processing module 702 is used to determine an activity push strategy for a corresponding link based on the characteristic data, where the corresponding link is one of the following: a credit card application binding link, a credit card application active customer maintenance link, a credit card application customer churn warning link, and a credit card application churn customer recall link;

[0320] The push module 703 is configured to push corresponding activity information to target client terminals according to corresponding activity push strategies.

[0321] In a possible implementation, the determining module 701 is specifically configured to:

[0322] The type of the target customer is determined based on at least one of whether the target customer has bound a corresponding credit card in the credit card application and whether the target customer has started the credit card application within a preset time period.

[0323] In a possible implementation, the determining module 701 is specifically configured to:

[0324] If it is determined that the target customer is a customer who has not bound a card to a credit card application, then the corresponding feature data is determined to be the first card usage feature;

[0325] If the target customer is determined to be an active customer who has bound a credit card to an application, the corresponding feature data is determined to be the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature;

[0326] If it is determined that the target customer is a credit card application churn warning customer, the corresponding feature data is determined to be the third card usage feature, the second application behavior feature, and the second card usage and behavior composite feature;

[0327] If it is determined that the type of target customer is a credit card application churn customer, the corresponding feature data is determined to be the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature.

[0328] In a possible implementation, the processing module 702 is specifically configured to:

[0329] Based on the characteristic data, determine whether the target customer is a recommended customer in the credit card application binding process;

[0330] In response to the target customer being a push customer for a credit card application binding phase, an activity push strategy for the credit card application binding phase is determined.

[0331] In one possible implementation, determining whether the target customer is a push customer in the credit card application binding process based on the characteristic data includes:

[0332] Inputting the first card usage feature into a card binding propensity prediction model, training the card binding propensity prediction model until convergence, and using the card binding propensity prediction model that has been trained to convergence to determine a card binding propensity score corresponding to the target customer in the card binding process of the credit card application. The card binding propensity prediction model that has been trained to convergence is located in a card binding module of the credit card application.

[0333] Based on the card binding propensity score, determine whether the target customer is a recommended customer for the credit card application card binding process.

[0334] In a possible implementation, before inputting the first card usage feature into the card binding tendency prediction model, the processing module 702 is further configured to:

[0335] Based on the churn customer recall module, obtain the churn recall scores corresponding to the credit card application churn customers among the target customers;

[0336] Based on the churn recall score, determine the non-churn recall intention customers and delete them from the target customers.

[0337] In a possible implementation, the processing module 702 is specifically configured to:

[0338] Determine the card binding tendency level based on the card binding tendency score;

[0339] In response to the card binding tendency level being the preset information push level, determining the target customer as a push customer for the credit card application program in the card binding process;

[0340] In response to the card binding tendency level not being the preset information push level, it is determined that the target customer is not a push customer in the card binding link of the credit card application.

[0341] In a possible implementation, the processing module 702 is specifically configured to:

[0342] Determine the corresponding promotion channel based on the target customers' card binding tendency level;

[0343] Obtain customer preference tags based on active customer maintenance in credit card applications;

[0344] Based on the corresponding activity push channels and customer preference tags, determine the activity push strategy for the corresponding card application binding process.

[0345] Push corresponding activity information to target customer terminals according to the corresponding activity push strategy, including:

[0346] Push activity information that matches the target customer's preference tags to the target customer terminal based on the corresponding activity push channel.

[0347] In a possible implementation, the processing module 702 is specifically configured to:

[0348] Determining at least one of a target customer's credit card application active behavior tag and a customer preference tag based on the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature;

[0349] Based on the credit card application binding process, obtain the target customer's activity push channel;

[0350] Determine a corresponding activity push strategy based on at least one of the target customer's credit card application active behavior tag, the customer preference tag, and / or the activity push channel;

[0351] Push corresponding activity information to target customer terminals according to the corresponding activity push strategy, including:

[0352] Pushing activity information to the target customer terminal via the credit card application based on at least one of the target customer's credit card application active behavior tag and customer preference tag;

[0353] And / or, at least one of the target customer's credit card application active behavior tag and the customer preference tag is used to push activity information to the target customer terminal through the target customer's activity push channel.

[0354] In a possible implementation, the processing module 702 is specifically configured to:

[0355] Based on the active customer maintenance phase of the credit card application, obtain the active behavior label of the credit card application;

[0356] Inputting the third card usage feature, the second application behavior feature, the second card usage and behavior composite feature, and the credit card application active behavior label into a churn warning model, training the churn warning model until convergence, and using the churn warning model that has been trained to convergence to determine a churn warning score corresponding to the target customer in the credit card application customer churn warning phase. The churn warning model that has been trained to convergence is located in a customer churn warning module.

[0357] Based on the third card usage feature, obtain the target customer's card churn category label. The target customer's card usage flow type label is used to indicate the probability that the target customer will not use the corresponding credit card;

[0358] Determine the churn customer group to which the target customer belongs based on their churn warning score and card usage churn category label in the credit card application's customer churn warning process;

[0359] Based on the churn customer group to which the target customer belongs, determine the corresponding activity push strategy in the customer churn warning link of the credit card application.

[0360] In a possible implementation, the processing module 702 is specifically configured to:

[0361] If it is determined that the target customer belongs to the first-level churn customer group, the activity push strategy corresponding to the customer churn warning phase of the credit card application is determined to be the first-level churn push strategy;

[0362] If it is determined that the target customer belongs to the second-level churn customer group, the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the second-level churn push strategy;

[0363] If it is determined that the target customer belongs to the third-level churn customer group, then the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the third-level churn push strategy;

[0364] If it is determined that the target customer belongs to the fourth-level churn customer group, then the corresponding activity push strategy in the credit card application's customer churn warning phase is determined to be the fourth-level churn push strategy;

[0365] In a possible implementation, the processing module 702 is specifically configured to:

[0366] If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to the first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to the second threshold, then the target customer belongs to the first-level churn customer group;

[0367] If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to the first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than the second threshold, then the target customer belongs to the second-level churn customer group;

[0368] If the target customer's churn warning score in the credit card application's customer churn warning phase is less than the first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to the second threshold, then the target customer belongs to the third-level churn customer group;

[0369] If the target customer's churn warning score corresponding to the credit card application's customer churn warning link is less than the first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than the second threshold, then the target customer belongs to the fourth-level churn customer group.

[0370] In a possible implementation, the processing module 702 is specifically configured to:

[0371] inputting the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature into a churn recall model, training the churn recall model until convergence, and using the churn recall model that has been trained to convergence to determine a churn recall score corresponding to the target customer in the credit card application churn customer recall phase, wherein the churn recall model that has been trained to convergence is located in a churn customer recall module;

[0372] Obtain the target customer's churn warning score from the customer churn warning module;

[0373] Determine the recall customer group to which the target customer belongs based on their churn recall score and churn warning score in the credit card application's churn recall phase;

[0374] Obtaining the target customer's card transaction preference tag based on the fourth card usage characteristic;

[0375] Based on at least one of the recall customer group to which the target customer belongs and the card transaction preference tag, a corresponding activity push strategy for recalling churned customers in a credit card application is determined.

[0376] In one possible implementation, determining the recall customer group to which the target customer belongs based on the target customer's churn recall score and churn warning score corresponding to the churn recall phase of the credit card application includes:

[0377] If the target customer's corresponding churn recall score is greater than or equal to the third threshold, and the churn warning score is greater than or equal to the fourth threshold, the target customer's recall customer group is determined to be the first-level recall customer group;

[0378] If the churn recall score corresponding to the target customer is greater than or equal to the third threshold, and the churn warning score is less than the fourth threshold, the target customer is determined to belong to the second-level recall customer group;

[0379] If the churn recall score corresponding to the target customer is less than the third threshold, and the churn recall score corresponding to the target customer is greater than or equal to the fifth threshold, then the target customer is determined to belong to the third-level recall customer group;

[0380] If the churn recall score corresponding to the target customer is less than the fifth threshold, the target customer is determined to belong to the fourth-level recall customer group;

[0381] In a possible implementation, the processing module 702 is specifically configured to:

[0382] If it is determined that the churned customer group to which the target customer belongs is a first-level recall customer group, then the activity push strategy corresponding to the churned customer recall phase of the credit card application is determined to be a first-level recall push strategy, which includes pushing activity information to the target customer;

[0383] If it is determined that the target customer belongs to a second-level recall customer group, then the corresponding activity push strategy for the lapsed customer recall phase of the credit card application is determined to be a second-level recall push strategy. The second-level push strategy includes pushing activity information matching the card transaction preference tag to the target customer.

[0384] If it is determined that the target customer belongs to the third-level recall customer group, the corresponding activity push strategy for the churn customer recall phase of the credit card application is determined to be the third-level recall push strategy. The third-level recall push strategy includes pushing activity information matching the card transaction preference tag to the target customer.

[0385] If it is determined that the churned customer group to which the target customer belongs is the fourth-level recall customer group, then the activity push strategy corresponding to the churned customer recall link of the credit card application is determined to be the fourth-level recall push strategy.

[0386] The data processing device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail in this embodiment.

[0387] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 8 As shown, the electronic device 800 provided in this embodiment includes: at least one processor 801 and a memory 802. Optionally, the device 800 also includes a communication component 803. The processor 801, the memory 802 and the communication component 803 are connected via a bus.

[0388] During the specific implementation process, at least one processor 801 executes the computer-executable instructions stored in the memory 802, so that the at least one processor 801 performs the above method.

[0389] The specific implementation process of the processor 801 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0390] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly implemented by a hardware processor or implemented by a combination of hardware and software modules in the processor.

[0391] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (NVM), such as at least one disk memory.

[0392] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be classified into address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0393] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0394] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0395] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0396] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0397] The division of units is merely a logical functional division; actual implementations may employ alternative divisions, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, any direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units, either through an interface, electrical, mechanical, or other means.

[0398] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0399] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0400] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0401] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0402] Finally, it should be noted that those skilled in the art will readily identify other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. The present invention is not limited to the precise structure described above and illustrated in the accompanying drawings, and various modifications and variations may be made without departing from the scope thereof. The scope of the present invention is limited solely by the appended claims.

Claims

1. A data processing method, characterized in that: The method includes: Determining a type of target customer to be processed, where the type of target customer is at least one of the following: a customer who has not bound a credit card application, a customer who has an active credit card application, a customer who has been warned of churn in a credit card application, or a customer who has churned in a credit card application; Determining corresponding characteristic data based on the type of the target customer, the characteristic data including at least one of the following: card usage characteristics, credit card application behavior characteristics, and card usage and behavior composite characteristics, wherein the card usage and behavior composite characteristics are determined based on the card usage characteristics and the credit card application behavior characteristics; Determining, based on the characteristic data, an activity push strategy for a corresponding link, wherein the corresponding link is one of the following: a credit card application binding link, a credit card application active customer maintenance link, a credit card application customer churn warning link, and a credit card application churn customer recall link; Push corresponding activity information to target client terminals according to the corresponding activity push strategy.

2. The method according to claim 1, characterized in that Determining the type of target customers to be processed includes: The type of the target customer is determined based on at least one of whether the target customer has bound a corresponding credit card in the credit card application and whether the target customer has started the credit card application within a preset time period.

3. The method according to claim 1, characterized in that The determining corresponding characteristic data according to the type of the target customer includes: If it is determined that the target customer is a customer whose credit card application has not been bound to a card, then the corresponding feature data is determined to be a first card usage feature; If it is determined that the target customer is an active customer who has bound a credit card to an application, the corresponding feature data is determined to be the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature; If it is determined that the target customer is a credit card application churn warning customer, then the corresponding feature data is determined to be the third card usage feature, the second application behavior feature, and the second card usage and behavior composite feature; If it is determined that the type of the target customer is a credit card application churn customer, the corresponding feature data is determined to be the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature.

4. The method according to claim 3, characterized in that The target customer type is a customer who has not bound a credit card to an application program. The activity push strategy for the corresponding link is determined based on the characteristic data, including: Determining, based on the characteristic data, whether the target customer is a push customer in the credit card application binding phase; In response to the target customer being a push customer in the credit card application binding phase, an activity push strategy for the credit card application binding phase is determined.

5. The method according to claim 4, characterized in that Determining, based on the characteristic data, whether the target customer is a push customer in the credit card application binding phase includes: inputting the first card usage feature into a card binding propensity prediction model, training the card binding propensity prediction model until convergence, and using the card binding propensity prediction model that has been trained to convergence to determine a card binding propensity score corresponding to the target customer in a credit card application binding process, wherein the card binding propensity prediction model that has been trained to convergence is located in a card binding module of the credit card application; Based on the card binding propensity score, it is determined whether the target customer is a recommended customer in the card binding link of the credit card application.

6. The method according to claim 5, characterized in that Before inputting the first card usage feature into the card binding tendency prediction model, the method further includes: Based on the churned customer recall module, obtaining churn recall scores corresponding to churned credit card application customers among the target customers; Based on the churn recall score, non-churn recall-willing customers are determined, and the non-churn recall-willing customers are deleted from the target customers.

7. The method according to claim 5, characterized in that The determining, based on the card binding propensity score, whether the target customer is a recommended customer for the credit card application program in the card binding process includes: Determining a card binding tendency level based on the card binding tendency score; In response to the card binding tendency level being the preset information push level, determining the target customer as a push customer for the credit card application program card binding step; In response to the card binding tendency level not being the preset information push level, it is determined that the target customer is not a push customer in the card binding link of the credit card application.

8. The method according to claim 4, characterized in that The activity push strategy for determining the card application binding process includes: Determine the corresponding promotion channel based on the target customer's card binding tendency level; obtaining a customer preference tag based on an active customer maintenance phase of the credit card application; Determine the activity push strategy for the corresponding card application binding process based on the corresponding activity push channel and customer preference tag; Pushing the corresponding activity information to the target client terminal according to the corresponding activity push strategy includes: The activity information matching the preference tag of the target customer is pushed to the target customer terminal based on the corresponding activity push channel.

9. The method according to claim 3, characterized in that The target customer type is an active customer who has bound a credit card application. The activity push strategy for the corresponding link is determined based on the characteristic data, including: determining at least one of a credit card application active behavior tag and a customer preference tag of the target customer based on the second card usage feature, the first application behavior feature, and the first card usage and behavior composite feature; Based on the credit card application binding process, obtaining the target customer's activity push channel; Determining a corresponding activity push strategy based on at least one of the target customer's credit card application activity behavior tag, the customer preference tag, and / or the activity push channel; The pushing of the corresponding activity information to the target client terminal according to the corresponding activity push strategy includes: Pushing activity information to the target customer terminal via the credit card application according to at least one of the target customer's credit card application active behavior tag and the customer preference tag; And / or, at least one of the target customer's credit card application active behavior tag and customer preference tag is used to push activity information to the target customer terminal through the target customer's activity push channel.

10. The method according to claim 3, characterized in that The target customer type is a credit card application churn warning customer, and the activity push strategy for the corresponding link is determined based on the characteristic data, including: Based on the credit card application active customer maintenance phase, obtaining a credit card application active behavior tag; Inputting the third card usage feature, the second application behavior feature, the second card usage and behavior composite feature, and the credit card application active behavior label into a churn warning model, training the churn warning model until convergence, and using the churn warning model that has been trained to convergence to determine a churn warning score corresponding to the target customer in the credit card application customer churn warning link, wherein the churn warning model that has been trained to convergence is located in a customer churn warning module; obtaining a card churn category label of the target customer based on the third card usage feature, wherein the card churn category label of the target customer is used to indicate a probability that the target customer will not use the corresponding credit card; Determining the churn customer group to which the target customer belongs based on the churn warning score corresponding to the target customer in the credit card application's customer churn warning link and the card usage churn category label; Based on the churn customer group to which the target customer belongs, an activity push strategy corresponding to the customer churn warning link of the credit card application is determined.

11. The method according to claim 10, characterized in that The determining of the activity push strategy corresponding to the credit card application customer churn warning link based on the churn customer group to which the target customer belongs includes: If it is determined that the churn customer group to which the target customer belongs is a first-level churn customer group, then determining that the activity push strategy corresponding to the customer churn warning link of the credit card application is a first-level churn push strategy; If it is determined that the churn customer group to which the target customer belongs is the second-level churn customer group, then determining that the activity push strategy corresponding to the customer churn warning link of the credit card application is the second-level churn push strategy; If it is determined that the churn customer group to which the target customer belongs is the third-level churn customer group, then the activity push strategy corresponding to the customer churn warning link of the credit card application is determined to be the third-level churn push strategy; If it is determined that the churn customer group to which the target customer belongs is a fourth-level churn customer group, then the activity push strategy corresponding to the customer churn warning link in the credit card application is determined to be a fourth-level churn push strategy.

12. The method according to claim 10, characterized in that Determining the churn customer group to which the target customer belongs based on the churn warning score corresponding to the target customer in the credit card application customer churn warning link and the card usage churn category label, including: If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to a first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to a second threshold, then the target customer belongs to a first-level churn customer group; If the target customer's churn warning score in the credit card application's customer churn warning phase is greater than or equal to a first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than a second threshold, then the target customer belongs to a second-level churn customer group; If the target customer's churn warning score in the credit card application's customer churn warning phase is less than a first threshold, and the probability of not using a credit card corresponding to the card churn category label is greater than or equal to a second threshold, then the target customer belongs to a third-level churn customer group; If the target customer's churn warning score corresponding to the credit card application's customer churn warning link is less than a first threshold, and the probability of not using a credit card corresponding to the card churn category label is less than a second threshold, then the churn customer group to which the target customer belongs is a fourth-level churn customer group.

13. The method according to claim 3, characterized in that The target customer type is a credit card application churn customer, and determining the activity push strategy for the corresponding link based on the characteristic data includes: Inputting the fourth card usage feature, the third application behavior feature, and the third card usage and behavior composite feature into a churn recall model, training the churn recall model until convergence, and using the churn recall model that has been trained to convergence to determine a churn recall score corresponding to the target customer in the credit card application churn customer recall phase, wherein the churn recall model that has been trained to convergence is located in a churn customer recall module; Obtaining a churn warning score for the target customer from a customer churn warning module; Determining the recall customer group to which the target customer belongs based on the churn recall score corresponding to the target customer in the churn customer recall phase of the credit card application and the churn warning score; acquiring a card transaction preference tag of the target customer based on the fourth card usage characteristic; Based on at least one of the recall customer group to which the target customer belongs and the card transaction preference tag, an activity push strategy corresponding to the churn customer recall phase of the credit card application is determined.

14. The method according to claim 13, characterized in that Determining the recalled customer group to which the target customer belongs based on the churn recall score corresponding to the target customer in the churn customer recall phase of the credit card application and the churn warning score includes: If the churn recall score corresponding to the target customer is greater than or equal to the third threshold, and the churn warning score is greater than or equal to the fourth threshold, then the recalled customer group to which the target customer belongs is determined to be the first-level recalled customer group; If the churn recall score corresponding to the target customer is greater than or equal to the third threshold, and the churn warning score is less than the fourth threshold, then the recalled customer group to which the target customer belongs is determined to be the second-level recalled customer group; If the churn recall score corresponding to the target customer is less than the third threshold, and the churn recall score corresponding to the target customer is greater than or equal to the fifth threshold, then the recalled customer group to which the target customer belongs is determined to be the third-level recalled customer group; If the churn recall score corresponding to the target customer is less than the fifth threshold, it is determined that the recall customer group to which the target customer belongs is the fourth-level recall customer group.

15. The method according to claim 13, characterized in that The determining of the activity push strategy corresponding to the churned customer recall phase of the credit card application based on at least one of the recall customer group to which the target customer belongs and the card transaction preference tag includes: If it is determined that the churned customer group to which the target customer belongs is a first-level recall customer group, then determining that the activity push strategy corresponding to the churned customer recall phase of the credit card application is a first-level recall push strategy, wherein the first-level push strategy includes pushing activity information to the target customer; If it is determined that the churned customer group to which the target customer belongs is a second-level recall customer group, then determining that the activity push strategy corresponding to the churned customer recall phase of the credit card application is a second-level recall push strategy, wherein the second-level push strategy includes pushing activity information matching the card transaction preference tag to the target customer; If it is determined that the churned customer group to which the target customer belongs is a third-level recall customer group, then determining that the activity push strategy corresponding to the churned customer recall phase of the credit card application is a third-level recall push strategy, wherein the third-level recall push strategy includes pushing activity information matching the card transaction preference tag to the target customer; If it is determined that the lost customer group to which the target customer belongs is a fourth-level recall customer group, then the activity push strategy corresponding to the lost customer recall link of the credit card application is determined to be a fourth-level recall push strategy.

16. A data processing device, characterized in that: The device includes: a determination module, configured to determine a type of target customer to be processed, wherein the type of target customer is at least one of the following: a customer whose credit card application has not been bound to a card, a customer whose credit card application has been actively bound to a card, a customer whose credit card application has been warned of churn, or a customer whose credit card application has churned; The determination module is further configured to determine corresponding characteristic data based on the type of the target customer, the characteristic data comprising at least one of the following: card usage characteristics, credit card application behavior characteristics, and card usage and behavior composite characteristics, wherein the card usage and behavior composite characteristics are determined based on the card usage characteristics and the credit card application behavior characteristics; a processing module, configured to determine, based on the characteristic data, an activity push strategy for a corresponding link, wherein the corresponding link is one of the following: a credit card application binding link, a credit card application active customer maintenance link, a credit card application customer churn warning link, and a credit card application churn customer recall link; The push module is used to push corresponding activity information to the target client terminal according to the corresponding activity push strategy.

17. An electronic device, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 15.

18. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 15 when executed by a processor.

19. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 15 when executed by a processor.

Images