Enterprise digital customer relationship management method and system

Abstract

The invention relates to the technical field of customer relationship management, in particular to an enterprise digital customer relationship management method and system. The method comprises the following steps: acquiring attribute data and behavior data of a target user; analyzing the attribute data and the behavior data, and generating a plurality of independent portrait modules; combining the independent portrait modules through a genetic algorithm, and generating an initial user portrait containing a multi-branch candidate label set; and real-time interaction data of the target user is obtained, pruning optimization is performed on the multi-branch candidate label set in the initial user portrait according to the real-time interaction data, and an accurate user portrait is obtained to perform customer relationship management. And more comprehensive analysis of user demands is facilitated, and personalized service and marketing strategies are formulated. By acquiring and analyzing the real-time interaction data, the initial user portrait is updated in time, no relevant labels are removed, the label most relevant to the latest behavior of the user is reserved, and the accuracy and timeliness of the accurate user portrait are improved.

Description

Technical Field

[0001] This application relates to the technical field of customer relationship management, and particularly to an enterprise digital customer relationship management method and system. Background Art

[0002] In today's business environment, enterprise digital customer relationship management has become one of the core strategies for building long-term customer relationships. With data-driven decision-making gradually becoming the mainstream, the importance of accurate and comprehensive user portraits in customer management has become increasingly prominent.

[0003] Traditional user portrait construction methods rely on the collection and analysis of a large amount of data, and then form user groups through clustering analysis, and infer customer needs based on this. However, how to construct a high-precision user portrait has become an important challenge for enterprises. Summary of the Invention

[0004] This application provides an enterprise digital customer relationship management method and system to solve the above problems.

[0005] In a first aspect, this application provides an enterprise digital customer relationship management method, and the method includes: Obtain the attribute data and behavior data of the target user; Analyze the attribute data and the behavior data to generate multiple independent portrait modules; Combine the independent portrait modules through a genetic algorithm to generate an initial user portrait including a multi-branch candidate tag set; Obtain the real-time interaction data of the target user, and according to the real-time interaction data, prune and optimize the multi-branch candidate tag set in the initial user portrait to obtain an accurate user portrait for customer relationship management.

[0006] Through this solution, obtaining the attribute data and behavior data of the target user helps to provide the basic information for constructing the initial user portrait, including the user's personal information and behavior records, and at the same time helps to identify the user's needs and preferences, and formulate personalized service and marketing strategies. By analyzing the attribute data, the basic characteristics of the user, such as age, gender, geographical location, etc., are determined, which helps to identify the basic attributes and background information of the user. At the same time, by analyzing the behavior data, the unique characteristics of the user, such as click behavior, purchase behavior, stay duration, etc., are determined, which helps to identify the behavior patterns and preferences of the user. The genetic algorithm optimizes the process of combining independent portrait modules to generate an initial user portrait with a multi-branch candidate tag set, which reflects the diversity and complexity of the user, helps to analyze the user's needs more comprehensively, and formulate personalized service and marketing strategies. By obtaining and analyzing real-time interaction data, the initial user portrait is updated in a timely manner, removing the tags that are no longer relevant and retaining the tags that are most relevant to the user's latest behavior, which helps to improve the accuracy and timeliness of the accurate user portrait, so as to better meet the user's needs and improve user satisfaction and loyalty.

[0007] Optionally, analyzing the attribute data and the behavior data to generate multiple independent portrait modules includes: Analyzing the attribute data to determine the basic characteristics; Analyzing the behavior data to determine the unique characteristics; Predicting the implicit characteristics according to the basic characteristics and the unique characteristics; Determining the basic portrait module according to the basic characteristics; Determining the unique portrait module according to the unique characteristics; Determining the implicit portrait module according to the implicit characteristics; Taking the basic portrait module, the unique portrait module, and the implicit portrait module as multiple independent portrait modules.

[0008] Through this solution, identifying the basic characteristics helps to identify the basic attributes and background information of the user, so as to better locate and segment the user group. Determining the unique characteristics helps to identify the behavior patterns and preferences of the user, so as to better predict the user's needs and behaviors. By predicting the implicit characteristics of the user, it helps to more deeply identify the inner needs and potential behaviors of the user, so as to formulate more accurate marketing and service strategies. The basic portrait module is the foundation of user portrait construction, which helps to establish the basic outline and characteristics of the user. The unique portrait module is the key to user portrait construction, which helps to distinguish the differences and personalized needs between users. The implicit portrait module helps to discover the inner needs and potential behaviors of the user, so as to formulate more accurate marketing and service strategies. Taking the basic portrait module, the unique portrait module, and the implicit portrait module as multiple independent portrait modules helps to conduct multi-dimensional analysis and identification of the user, making the construction of the user portrait more flexible.

[0009] Optionally, the attribute data includes age, gender, and geographical location; predicting the implicit characteristics based on the basic characteristics and the unique characteristics includes: Analyze the behavior data to determine click behavior, purchase behavior, and dwell time; Analyze the age and gender to determine the general needs; Analyze the geographical location, the click behavior, the purchase behavior, and the dwell time to determine the potential needs; Predict the implicit characteristics based on the general needs and the potential needs.

[0010] Through this solution, by analyzing the behavior data, the user's interest points, purchase preferences, and engagement are identified. Click behavior reflects the user's focus and interest points, purchase behavior reveals the user's consumption habits and preferences, and dwell time evaluates the user's interest and engagement with the content. By analyzing age and gender, the general needs of different groups are identified. For example, young users are more inclined to innovative products and online shopping, while older users are more concerned about product quality and after-sales service. By analyzing data such as geographical location, click behavior, purchase behavior, and dwell time, the potential needs of users under different geographical locations and behavior data are identified. For example, users in different regions have higher interests in different product categories, or the browsing behavior of users at different time periods indicates upcoming purchase behavior. By predicting the implicit characteristics, the user's inner needs and potential behaviors are more deeply identified, which helps to formulate more precise marketing and service strategies. For example, according to the prediction results, products or services that the user is interested in are recommended to the user.

[0011] Optionally, the combination of the independent portrait modules through the genetic algorithm to generate an initial user portrait including a multi-branch candidate tag set includes: Determine the initial gene template according to the basic portrait module; Through the genetic algorithm, analyze the unique portrait modules to determine the co-occurrence probability of any two unique portrait modules; Determine the module connection between any two unique portrait modules according to the co-occurrence probability; Randomly cross-combine the unique portrait modules according to the module connection to generate mutant candidate tags; Determine a number of implicit tags according to the implicit portrait module; Analyze the mutant candidate tags and the number of implicit tags to determine the tag conflict rate; Adjust the initial gene template according to the tag conflict rate to generate an initial user portrait including a multi-branch candidate tag set.

[0012] Through this solution, the initial gene template provides a basic framework for genetic algorithm operations, helps to maintain the core features of the user profile, provides a starting point for the construction of the user profile, and ensures the consistency and stability of the user profile. The co-occurrence probability reflects the frequency of simultaneous occurrence of different unique profile modules. By analyzing the co-occurrence probability, the relationships between different unique profile modules are identified. The module connection reflects the relevance and interaction between different unique profile modules. By determining the module connection, the complexity and diversity of the user profile are constructed, making the user profile more comprehensive and accurate. Through the randomly cross-combined unique profile modules, new user profile mutation candidate tags are generated, thus enriching the diversity of the user profile. The implicit tags reflect the potential needs and preferences of users, such as potential purchase intentions, future behavior trends, etc. By determining the implicit tags, the inherent needs and potential behaviors of users are revealed, providing deeper insights. The label conflict rate reflects the inconsistency and contradiction degree between the mutation candidate tags and the implicit tags. By analyzing the label conflict rate, the conflicts and contradictions in the user profile are identified and resolved, improving the accuracy and consistency of the user profile. Adjusting the initial gene template helps to optimize the construction process of the user profile and reduce the label conflict rate. Generating an initial user profile with a multi-branch candidate tag set helps to analyze customer needs more comprehensively, explore different user profile combinations, and thus better meet customer needs.

[0013] Optionally, pruning and optimizing the multi-branch candidate tag set in the initial user profile according to the real-time interaction data to obtain an accurate user profile includes: Extract the interaction type, interaction frequency, and interaction depth from the real-time interaction data; Match the interaction type with the multi-branch candidate tags to determine the concentrated associated tags; Based on the interaction frequency and interaction depth, calculate the real-time confidence score of the concentrated associated tags; Compare the real-time confidence score with a preset dynamic threshold, and screen the concentrated associated tags according to the comparison result to obtain qualified tags; Generate the accurate user profile according to the qualified tags.

[0014] Through this solution, by obtaining the interaction type, interaction frequency, and interaction depth, and analyzing the real-time behavior data of users, it helps to timely adjust services and marketing strategies to adapt to the rapidly changing market demands. Through matching, determining the concentrated associated tags for real-time interaction with users helps to more accurately locate the needs of users and provide more personalized services. The real-time confidence score helps to evaluate the accuracy and relevance of the concentrated associated tags, and at the same time helps to identify which concentrated associated tags can better reflect the true needs of users. Through screening, removing the concentrated associated tags that are no longer associated with users and retaining the concentrated associated tags that accurately reflect the current state of users helps to maintain the accuracy and timeliness of the user portrait. The accurate user portrait reflects the latest behaviors and preferences of users, provides a basis for more accurate market positioning and personalized services, helps to improve customer satisfaction and loyalty, and enhances market competitiveness.

[0015] Optionally, predicting the implicit features according to the mass demands and the potential demands includes: Analyze the behavior data to determine a set of high-frequency behavior tags; According to the mass demands and the potential demands, construct a set of mass tags and a set of potential tags; Calculate the first set similarity and the second set similarity between the set of high-frequency behavior tags and the set of mass tags and the set of potential tags respectively; Compare the first set similarity and the second set similarity with a preset similarity threshold respectively. If any one of the first set similarity and the second set similarity is less than the preset similarity threshold, extract the time-series behavior sequence from the behavior data; According to the time-series behavior sequence, eliminate the overlapping behaviors in the set of mass tags and the set of potential tags; Input the set of mass tags after elimination and the set of potential tags after elimination into a preset demand prediction model to predict the implicit features.

[0016] Through this solution, the high-frequency behavior tag set helps to identify users' daily behavior habits and preferences, thus better positioning and segmenting user groups. Constructing the mass tag set and the potential tag set according to the mass demand and potential demand helps to conduct market segmentation and personalized marketing. By calculating the first set similarity and the second set similarity between the high-frequency behavior tag set and the mass tag set, and the potential tag set, it is determined whether it is necessary to further analyze the user's behavior data. By comparing the first set similarity and the second set similarity with the preset similarity threshold, it is judged whether the user's behavior data is consistent with the preset demand characteristics. If the first set similarity and the second set similarity are lower than the preset similarity threshold, it indicates that the user's behavior data is inconsistent with the preset demand characteristics, and it is necessary to further analyze the user's time-series behavior sequence. Extracting the time-series behavior sequence in the behavior data helps to discover the trends and changes in user behavior, thus better identifying the user's needs and behavior data. Eliminating the overlapping behaviors in the mass tag set and the potential tag set helps to simplify the preset demand prediction model and improve the efficiency and accuracy of the preset demand prediction model. Inputting the eliminated mass tag set and potential tag set into the preset demand prediction model to predict the user's implicit characteristics helps to more comprehensively identify the user, formulate personalized service and marketing strategies, and improve user satisfaction.

[0017] Optionally, the analyzing the behavior data to determine unique characteristics includes: Extracting the event type and behavior timestamp in the behavior data; According to the event type, mining high-frequency behavior combinations through the FP-growth algorithm to determine behavior pattern primitives; Analyzing the behavior timestamp and calculating the time interval distribution of adjacent behaviors; According to the time interval distribution, determining the behavior time characteristics; According to the behavior pattern primitives and the behavior time characteristics, determining unique characteristics.

[0018] Through this solution, by extracting event types and behavior timestamps to identify users' interaction behaviors and time patterns, it helps to better analyze users' behavior habits. Behavioral pattern primitives reveal common interaction patterns of users, which helps to identify users' behavior preferences. Analyze the behavior timestamps, calculate the time interval distribution between adjacent behaviors, and analyze the time rules of users' behaviors. The time interval distribution reveals the time patterns of users' behaviors, which helps to identify users' behavior cycles and active time periods. According to the time interval distribution, determine users' behavioral time characteristics, such as active periods, behavior cycles, etc., which helps to analyze users' behavior patterns and time preferences, providing a basis for personalized services. Combining behavioral pattern primitives and behavioral time characteristics, determine users' unique features, reflecting users' personalized behavior patterns, which helps to more deeply identify users' personalized needs and provide support for customized services and marketing strategies.

[0019] Optionally, the adjusting the initial gene template according to the tag conflict rate includes: Determine the conflict type according to the mutation candidate tags and the several recessive tags; Analyze the conflict type to determine the conflict pattern; Determine the template adjustment scheme according to the conflict pattern and the tag conflict rate; Adjust the initial gene template according to the template adjustment scheme.

[0020] Through this solution, by analyzing the conflicts between mutation candidate tags and several recessive tags, identify the inconsistencies between mutation candidate tags and several recessive tags, providing a basis for conflict pattern analysis. By analyzing the conflict types, identify the conflict patterns, reflecting the mutual relationship between mutation candidate tags and several recessive tags, providing guidance for the template adjustment scheme. According to the conflict pattern and the tag conflict rate, formulate the template adjustment scheme to improve the accuracy and consistency of the user portrait. According to the template adjustment scheme, adjust the initial gene template, thereby optimizing the construction process of the user portrait.

[0021] Optionally, the method further includes: Continuously monitor the continuous interaction data of the target user; Analyze the continuous interaction data to determine the updated data; According to the updated data, dynamically expand the multi-branch candidate tag set of the accurate user portrait, and perform pruning optimization on the expanded multi-branch candidate tag set.

[0022] Through this solution, by continuously monitoring the user's continuous interaction data and analyzing the user's behavior patterns in real time, it helps to adjust the service strategy in a timely manner to meet the rapidly changing market demands. By analyzing the user's continuous interaction data and identifying the user portrait data that needs to be updated, it helps to adjust the user portrait in a timely manner to reflect the user's latest behaviors and preferences, and improve the accuracy and timeliness of the user portrait. According to the updated data, dynamically expanding the multi-branch candidate tag set of the precise user portrait helps to identify the diverse needs and behaviors of users, and provides support for personalized services and marketing strategies. Using the pruning process in the genetic algorithm to optimize the expanded multi-branch candidate tag set and removing redundant or irrelevant candidate tags helps to simplify the user portrait, reduce redundant information, and improve the efficiency and practicality of the user portrait.

[0023] In a second aspect, the present application provides an enterprise digital customer relationship management system, which includes: A data acquisition module for acquiring the attribute data and behavior data of the target user; A data analysis module for analyzing the attribute data and the behavior data to generate multiple independent portrait modules; An initial portrait generation module for combining the independent portrait modules through a genetic algorithm to generate an initial user portrait including a multi-branch candidate tag set; A portrait optimization module for acquiring the real-time interaction data of the target user, and pruning and optimizing the multi-branch candidate tag set in the initial user portrait according to the real-time interaction data to obtain a precise user portrait for customer relationship management.

[0024] Optionally, when the data analysis module analyzes the attribute data and the behavior data to generate multiple independent portrait modules, it is used for: Analyzing the attribute data to determine the basic characteristics; Analyzing the behavior data to determine the unique characteristics; Predicting the implicit characteristics according to the basic characteristics and the unique characteristics; Determining the basic portrait module according to the basic characteristics; Determining the unique portrait module according to the unique characteristics; Determining the implicit portrait module according to the implicit characteristics; Taking the basic portrait module, the unique portrait module, and the implicit portrait module as multiple independent portrait modules.

[0025] Optionally, the attribute data includes age, gender, and geographical location; when the data analysis module predicts the implicit characteristics according to the basic characteristics and the unique characteristics, it is used for: Analyzing the behavior data to determine the click behavior, purchase behavior, and stay duration; Analyze the age and gender to determine the general public's needs; Analyze the geographical location, click behavior, purchase behavior, and stay duration to determine potential needs; Predict implicit characteristics based on the general public's needs and the potential needs.

[0026] Optionally, when the initial portrait generation module combines the independent portrait modules through a genetic algorithm to generate an initial user portrait containing a multi-branch candidate tag set, it is used for: Determine an initial gene template according to the basic portrait module; Analyze the unique portrait modules through a genetic algorithm to determine the co-occurrence probability of any two unique portrait modules; Determine the module connection between any two unique portrait modules according to the co-occurrence probability; Randomly cross-combine the unique portrait modules according to the module connection to generate mutant candidate tags; Determine a number of implicit tags according to the implicit portrait module; Analyze the mutant candidate tags and the number of implicit tags to determine the tag conflict rate; Adjust the initial gene template according to the tag conflict rate to generate an initial user portrait containing a multi-branch candidate tag set.

[0027] Optionally, when the portrait optimization module performs pruning optimization on the multi-branch candidate tag set in the initial user portrait according to the real-time interaction data to obtain an accurate user portrait, it is used for: Extract the interaction type, interaction frequency, and interaction depth from the real-time interaction data; Match the interaction type with the multi-branch candidate tags to determine the concentrated associated tags; Calculate the real-time confidence score of the concentrated associated tags based on the interaction frequency and interaction depth; Compare the real-time confidence score with a preset dynamic threshold, and screen the concentrated associated tags according to the comparison result to obtain qualified tags; Generate the accurate user portrait according to the qualified tags.

[0028] Optionally, when the data analysis module predicts implicit characteristics based on the general public's needs and the potential needs, it is used for: Analyze the behavior data to determine a set of high-frequency behavior tags; Construct a general public tag set and a potential tag set according to the general public's needs and the potential needs; Calculate the first set similarity and the second set similarity between the set of high-frequency behavior tags and the set of public tags and the set of potential tags respectively; Compare the first set similarity and the second set similarity with a preset similarity threshold respectively. If any one of the first set similarity and the second set similarity is less than the preset similarity threshold, extract the sequential behavior sequence in the behavior data; According to the sequential behavior sequence, eliminate the overlapping behaviors in the set of public tags and the set of potential tags; Input the set of public tags after elimination and the set of potential tags after elimination into a preset demand prediction model to predict the implicit characteristics.

[0029] Optionally, when the data analysis module analyzes the behavior data to determine the unique characteristics, it is used for: Extract the event type and behavior timestamp in the behavior data; According to the event type, mine high-frequency behavior combinations through the FP-growth algorithm to determine the behavior pattern primitives; Analyze the behavior timestamp and calculate the time interval distribution of adjacent behaviors; According to the time interval distribution, determine the behavior time characteristics; According to the behavior pattern primitives and the behavior time characteristics, determine the unique characteristics.

[0030] Optionally, when the initial portrait generation module adjusts the initial gene template according to the tag conflict rate, it is used for: Determine the conflict type according to the mutant candidate tags and the several implicit tags; Analyze the conflict type to determine the conflict pattern; Determine the template adjustment plan according to the conflict pattern and the tag conflict rate; Adjust the initial gene template according to the template adjustment plan.

[0031] Optionally, the enterprise digital customer relationship management system further includes a portrait expansion module, which is used for: Continuously monitor the continuous interaction data of the target user; Analyze the continuous interaction data to determine the updated data; According to the updated data, dynamically expand the multi-branch candidate tag set of the accurate user portrait and perform pruning optimization on the expanded multi-branch candidate tag set. Description of the Drawings

[0032] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0033] Figure 1 A schematic diagram of an application scenario provided by an embodiment of the present application; Figure 2 A flowchart of a method for enterprise digital customer relationship management provided by an embodiment of the present application; Figure 3 A schematic diagram of the structure of an enterprise digital customer relationship management system provided by an embodiment of the present application. Detailed implementation manners

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0035] In addition, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after, unless otherwise specified.

[0036] The following will further describe the embodiments of the present application in detail with reference to the accompanying drawings of the specification.

[0037] Traditional user portrait construction methods rely on the collection and analysis of a large amount of data, and then form user groups through clustering analysis, and infer customer needs based on this. However, how to construct a high-precision user portrait has become an important challenge for enterprises.

[0038] Based on this, the present application provides an enterprise digital customer relationship management method and system, which obtain the attribute data and behavior data of target users; analyze the attribute data and behavior data to generate multiple independent portrait modules; combine the independent portrait modules through a genetic algorithm to generate an initial user portrait containing a multi-branch candidate tag set; obtain the real-time interaction data of the target users, and prune and optimize the multi-branch candidate tag set in the initial user portrait according to the real-time interaction data to obtain an accurate user portrait for customer relationship management. Obtaining the attribute data and behavior data of target users helps to provide the basic information for constructing the initial user portrait, including the personal information and behavior records of users, and at the same time helps to identify user needs and preferences and formulate personalized services and marketing strategies. By analyzing the attribute data, the basic characteristics of users, such as age, gender, geographical location, etc., are determined, which helps to identify the basic attributes and background information of users. At the same time, by analyzing the behavior data, the unique characteristics of users, such as click behavior, purchase behavior, stay duration, etc., are determined, which helps to identify the behavior patterns and preferences of users. The process of combining independent portrait modules optimized by a genetic algorithm generates an initial user portrait with a multi-branch candidate tag set, reflecting the diversity and complexity of users, which helps to analyze user needs more comprehensively and formulate personalized services and marketing strategies. By obtaining and analyzing real-time interaction data, the initial user portrait is updated in a timely manner, removing the tags that are no longer relevant and retaining the tags that are most relevant to the latest behavior of users, which helps to improve the accuracy and timeliness of the accurate user portrait, so as to better meet user needs and improve user satisfaction and loyalty.

[0039] Figure 1 FIG. is a schematic diagram of an application scenario provided by the present application. When performing customer relationship management, the method provided by the present application is applied.

[0040] Specifically, the method provided by the present application is applied to any server. The server interacts with the internal database to obtain the attribute data and behavior data of the target users in the internal database, analyzes the attribute data to determine the basic characteristics of the users, and analyzes the behavior data to determine the unique characteristics of the users. The process of combining independent portrait modules optimized by a genetic algorithm generates an initial user portrait with a multi-branch candidate tag set. Through the acquisition and analysis of real-time interaction data, the multi-branch candidate tag set in the initial user portrait is pruned and optimized to obtain an accurate user portrait and store it in the internal database for customer relationship management. The specific implementation manner can refer to the following embodiments.

[0041] Figure 2 FIG. is a flowchart of an enterprise digital customer relationship management method provided by an embodiment of the present application. The method of this embodiment can be applied to the server in the above scenario. As Figure 2 shown, the method includes: S201. Obtain the attribute data and behavior data of the target user; The target user can be an individual or a group who hopes to identify the needs, behaviors, and preferences of the target user and establish and maintain relationships with the user.

[0042] Attribute data can be data such as age, gender, geographical location, occupation, educational background, etc. related to the target user.

[0043] Behavior data can be dynamic behavior data such as the click behavior, purchase behavior, browsing behavior, search behavior, interaction behavior, etc. of the target user.

[0044] Specifically, collect the attribute data of users through means such as registration forms, questionnaires, and social media data scraping. Collect the behavior data of users on products, websites, or applications, such as click behavior, browsing paths, purchase history, search records, interaction behavior, etc., through means such as website analysis tools, in-app tracking, and third-party data providers.

[0045] S202. Analyze the attribute data and behavior data to generate multiple independent portrait modules; The independent portrait module can be to decompose the user portrait into multiple independent modules, and each module represents a certain aspect of the user characteristics, such as the basic portrait module, unique portrait module, implicit portrait module, etc.

[0046] Specifically, perform preprocessing on the collected attribute data, such as removing missing values, handling outliers, and unifying data formats. Extract the basic characteristics of users from the processed attribute data. Perform preprocessing such as cleaning and standardization on the collected behavior data. Use data mining techniques to extract the unique characteristics of users from the processed behavior data. Generate multiple independent portrait modules based on the basic characteristics and unique characteristics.

[0047] S203. Combine the independent portrait modules through a genetic algorithm to generate an initial user portrait containing a multi-branch candidate tag set; The genetic algorithm can be an optimization algorithm that simulates the process of natural selection and heredity, and is an algorithm that searches and combines iteratively to find the optimal solution or an approximate optimal solution.

[0048] The multi-branch candidate tag set can be a multi-branch candidate tag set generated by optimizing the module combination through a genetic algorithm, and each tag set contains multiple branches, representing different attribute data and behavior data of the user.

[0049] The initial user portrait can be a preliminary user portrait generated by combining the independent portrait modules through a genetic algorithm.

[0050] Specifically, the construction of user portraits is a core part of customer management. In the existing technology, the basic process for constructing user portraits is: collection, processing, analysis (clustering), and tagging. This method requires a large amount of data to achieve a certain level of accuracy in the portrait. There is also another method, which is to first map users to a template based on basic information, and then make the portrait more comprehensive through a growth approach. This method also requires a large amount of data to achieve a certain level of accuracy in the portrait.

[0051] In this embodiment, the independent portrait module is encoded as a "chromosome" in the genetic algorithm, that is, an initial portrait. Generate a certain number of initial portraits. Define a fitness function for evaluating the quality of each initial portrait. According to the evaluation results of the fitness function, use methods such as tournament selection to select the initial user portraits with higher fitness for reproduction. Perform random crossover on the selected initial user portraits to generate new accurate user portraits. Perform mutation operations on the accurate user portraits to simulate gene mutations. Add the new accurate user portraits to the initial user portraits and replace the user portraits with lower fitness to form new initial user portraits. The genetic algorithm continuously iteratively optimizes the module combination by simulating natural selection and genetic processes to generate an initial user portrait containing a multi-branch candidate tag set.

[0052] S204. Obtain the real-time interaction data of the target user, and according to the real-time interaction data, perform pruning optimization on the multi-branch candidate tag set in the initial user portrait to obtain an accurate user portrait for customer relationship management.

[0053] The real-time interaction data can be real-time interaction records such as click behavior, purchase behavior, browsing behavior, search behavior, and interaction behavior of the target user with products, services, platforms, etc.

[0054] The accurate user portrait can be the final user portrait obtained by performing pruning optimization on the multi-branch candidate tag set in the initial user portrait.

[0055] Customer relationship management can be a series of strategies, activities, and measures taken to establish, maintain, and develop relationships with users.

[0056] Specifically, use methods such as Web analysis tools, in-app event tracking, and social media monitoring to obtain real-time interaction data such as the interaction type, interaction frequency, and interaction depth of the target user. According to the real-time interaction data, perform pruning optimization on the multi-branch candidate tag set. Based on the real-time interaction data of the user, calculate the confidence score of each multi-branch candidate tag. Set a dynamic threshold for screening the confidence of the multi-branch candidate tags. Update the user portrait according to the selected tags to generate an accurate user portrait, thereby performing customer relationship management.

[0057] Through this solution, obtaining the attribute data and behavior data of the target user helps to provide the basic information for constructing the initial user portrait, including the user's personal information and behavior records, and at the same time helps to identify the user's needs and preferences, and formulate personalized service and marketing strategies. By analyzing the attribute data to determine the basic characteristics of the user, such as age, gender, geographical location, etc., it helps to identify the basic attributes and background information of the user. At the same time, by analyzing the behavior data to determine the unique characteristics of the user, such as click behavior, purchase behavior, stay duration, etc., it helps to identify the user's behavior patterns and preferences. The genetic algorithm optimizes the process of combining independent portrait modules to generate an initial user portrait with a multi-branch candidate label set, reflecting the diversity and complexity of users, which helps to analyze the user's needs more comprehensively and formulate personalized service and marketing strategies. By obtaining and analyzing real-time interaction data, the initial user portrait is updated in a timely manner, removing the no-longer relevant labels and retaining the labels most relevant to the user's latest behavior, which helps to improve the accuracy and timeliness of the accurate user portrait, thus better meeting the user's needs and improving user satisfaction and loyalty.

[0058] In some embodiments, analyze the attribute data to determine the basic characteristics; analyze the behavior data to determine the unique characteristics; predict the implicit characteristics according to the basic characteristics and unique characteristics; determine the basic portrait module according to the basic characteristics; determine the unique portrait module according to the unique characteristics; determine the implicit portrait module according to the implicit characteristics; and use the basic portrait module, the unique portrait module, and the implicit portrait module as multiple independent portrait modules.

[0059] The basic characteristics can be static attributes of the user such as age, gender, geographical location, education level, occupation, income level, etc.

[0060] The unique characteristics can be dynamic behavior characteristics of the user such as click behavior, purchase history, browsing record, search record, interaction behavior, etc.

[0061] The implicit characteristics can be characteristics of the user that are potential and not easily directly observable.

[0062] The basic portrait module can be the basic part of the user portrait.

[0063] The unique portrait module can be the unique part of the user portrait.

[0064] The implicit portrait module can be the potential part of the user portrait.

[0065] Specifically, statistically analyze the collected attribute data to identify the basic characteristics representing the user. Analyze the user's behavior data to identify the unique characteristics representing the user. Combine the basic characteristics and the unique characteristics, and use a preset demand prediction model to predict the implicit characteristics of the user. Create a basic portrait module based on the basic characteristics. Create a unique portrait module based on the unique characteristics. Create an implicit portrait module based on the implicit characteristics. Integrate the basic portrait module, the unique portrait module, and the implicit portrait module together to form multiple independent portrait modules.

[0066] Through this solution, identifying the basic characteristics helps to identify the basic attributes and background information of the user, thus better positioning and segmenting the user group. Determining the unique characteristics helps to identify the behavior patterns and preferences of the user, thus better predicting the user's needs and behaviors. Predicting the implicit characteristics of the user helps to more deeply identify the inner needs and potential behaviors of the user, thus formulating more precise marketing and service strategies. The basic portrait module is the foundation of user portrait construction, which helps to establish the basic profile and characteristics of the user. The unique portrait module is the key to user portrait construction, which helps to distinguish the differences and personalized needs between users. The implicit portrait module helps to discover the inner needs and potential behaviors of the user, thus formulating more precise marketing and service strategies. Taking the basic portrait module, the unique portrait module, and the implicit portrait module as multiple independent portrait modules helps to conduct multi-dimensional analysis and identification of the user, making the construction of the user portrait more flexible.

[0067] In some embodiments, analyze the behavior data to determine click behavior, purchase behavior, and dwell time; analyze age and gender to determine the general needs; analyze geographical location, click behavior, purchase behavior, and dwell time to determine potential needs; predict the implicit characteristics based on the general needs and potential needs.

[0068] The click behavior can be click operations such as clicking on links, buttons, pictures, etc. on a website, application, or advertisement.

[0069] The purchase behavior can be the behavior of a user purchasing goods or services.

[0070] The dwell time can be the length of time spent staying on a web page, application page, or content.

[0071] The general needs can be the needs or preferences commonly existing in a group.

[0072] The geographical location can be the physical location where the user is located.

[0073] The potential needs can be the needs that the user has not clearly expressed or has not been satisfied.

[0074] Specifically, based on the behavioral data, data mining techniques are used to analyze the user's click behavior, purchase behavior, and the duration of stay on different pages or content. Statistical analysis methods are used to analyze the user's age and gender to determine the general public's needs. Combining data such as the user's geographical location, click behavior, purchase behavior, and duration of stay, sequence pattern mining techniques are used for analysis to determine the user's potential needs. A preset demand prediction model is used to predict the user's implicit characteristics.

[0075] Through this solution, by analyzing the behavioral data, the user's interest points, purchase preferences, and engagement are identified. Click behavior reflects the user's focus and interest points, purchase behavior reveals the user's consumption habits and preferences, and the duration of stay assesses the user's interest and engagement with the content. By analyzing age and gender, the general public's needs of different groups are identified. For example, young users are more inclined to innovative products and online shopping, while older users are more concerned about product quality and after-sales service. By analyzing data such as geographical location, click behavior, purchase behavior, and duration of stay, the potential needs of users under different geographical locations and behavioral data are identified. For example, users in different regions have higher interest in different product categories, or the browsing behavior of users at different time periods indicates upcoming purchase behavior. By predicting implicit characteristics, the user's inner needs and potential behaviors can be identified more deeply, which helps to formulate more precise marketing and service strategies. For example, according to the prediction results, products or services that the user is interested in can be recommended to the user.

[0076] In some embodiments, according to the basic portrait module, an initial gene template is determined; through the genetic algorithm, the unique portrait modules are analyzed to determine the co-occurrence probability of any two unique portrait modules; according to the co-occurrence probability, the module connection between any two unique portrait modules is determined; according to the module connection, the unique portrait modules are randomly cross-combined to generate mutant candidate tags; according to the implicit portrait module, a number of implicit tags are determined; the mutant candidate tags and the number of implicit tags are analyzed to determine the tag conflict rate; according to the tag conflict rate, the initial gene template is adjusted to generate an initial user portrait containing a multi-branch candidate tag set.

[0077] The initial gene template can be a preset framework for representing the basic characteristics of the user portrait.

[0078] The co-occurrence probability can be the frequency of simultaneous occurrence of any two unique portrait modules.

[0079] The module connection can be the relevance and interaction between different unique portrait modules.

[0080] Random crossover can be in the genetic algorithm, where part of the genes of two unique portrait modules are exchanged to generate new unique portrait modules.

[0081] A mutation candidate tag can be a new tag generated by exchanging genes of the exclusive portrait module during the random crossover process.

[0082] A recessive tag can be a need that is not explicitly expressed or fully satisfied in the user portrait.

[0083] The tag conflict rate can be the degree of inconsistency and contradiction between the mutation candidate tags and the recessive tags in the user portrait.

[0084] Specifically, based on the basic portrait module, an initial gene template is constructed. The exclusive portrait modules are encoded as chromosomes in the genetic algorithm, that is, the user portrait. Through the iterative process of the genetic algorithm, the user portrait is continuously optimized to determine the co-occurrence probability of any two exclusive portrait modules. The co-occurrence probability is analyzed using the association rule learning algorithm to determine the association rules between different exclusive portrait modules. Based on the co-occurrence probability and the results of the association rule learning, the module connection between any two exclusive portrait modules is determined. According to the module connection, the exclusive portrait modules with higher fitness are selected for crossover combination. Crossover operations are performed on the selected exclusive portrait modules to exchange some features. Mutation operations are performed on the user portraits generated by the crossover to randomly change some features. Through the crossover and mutation operations, mutation candidate tags of new user portraits are generated. Based on the recessive portrait module, several recessive tags are defined. The mutation candidate tags and the recessive tags are aligned, and statistical methods are used to identify tag conflicts and calculate the tag conflict rate. According to the results of the tag conflict rate, the initial gene template is adjusted. Using the adjusted initial gene template, through the crossover and mutation operations of the genetic algorithm, an initial user portrait containing a multi-branch candidate tag set is generated.

[0085] Through this solution, the initial gene template provides a basic framework for genetic algorithm operations, helps maintain the core features of the user profile, provides a starting point for the construction of the user profile, and ensures the consistency and stability of the user profile. The co-occurrence probability reflects the frequency of different unique profile modules appearing simultaneously. By analyzing the co-occurrence probability, the relationships between different unique profile modules are identified. The module connection reflects the relevance and interaction between different unique profile modules. By determining the module connection, the complexity and diversity of the user profile are constructed, making the user profile more comprehensive and accurate. Through randomly cross-combined unique profile modules, new user profile mutation candidate tags are generated, thus enriching the diversity of the user profile. Implicit tags reflect the potential needs and preferences of users, such as potential purchase intentions, future behavior trends, etc. By determining the implicit tags, the internal needs and potential behaviors of users are revealed, providing deeper insights. The tag conflict rate reflects the inconsistency and contradiction between the mutation candidate tags and the implicit tags. By analyzing the tag conflict rate, the conflicts and contradictions in the user profile are identified and resolved, improving the accuracy and consistency of the user profile. Adjusting the initial gene template helps optimize the construction process of the user profile and reduce the tag conflict rate. Generating an initial user profile with a multi-branch candidate tag set helps analyze customer needs more comprehensively, explore different user profile combinations, and thus better meet customer needs.

[0086] In some embodiments, the interaction type, interaction frequency, and interaction depth in the real-time interaction data are extracted; the interaction type is matched with the multi-branch candidate tags to determine the centralized associated tags; based on the interaction frequency and interaction depth, the real-time confidence score of the centralized associated tags is calculated; the real-time confidence score is compared with a preset dynamic threshold, and the centralized associated tags are screened according to the comparison result to obtain qualified tags; and an accurate user profile is generated according to the qualified tags.

[0087] The interaction type can be the way users interact with products, services, or platforms.

[0088] The interaction frequency can be the number of times the user performs the interaction type within a certain period of time.

[0089] The interaction depth can be the degree of investment of the user in the interaction type.

[0090] The centralized associated tags can be the tags most relevant to the user's real-time interaction.

[0091] The real-time confidence score can be the confidence score of the centralized associated tags calculated based on the interaction frequency and interaction depth.

[0092] The preset dynamic threshold can be a dynamic threshold preset according to business requirements and behavior data for screening the centralized associated tags. It is stored in the server in advance and called when in use.

[0093] The compliance label can be a label selected based on the comparison result of the real-time confidence score and the preset dynamic threshold.

[0094] Specifically, use real-time data stream processing technology to capture and analyze the interaction type, interaction frequency, and interaction depth in real-time interaction data. Use a matching algorithm to match the interaction type in the real-time interaction data with a predefined multi-branch candidate label set. Based on the matching result, determine the centralized associated labels for the user's real-time interaction. According to the interaction frequency and interaction depth, use statistical analysis methods to calculate the real-time confidence score of the centralized associated labels. Set a dynamic threshold, which is adjusted according to the changes in the real-time interaction data. Compare the real-time confidence score with the preset dynamic threshold, and select the labels with scores higher than the preset dynamic threshold, which are the compliance labels. Update the user portrait based on the selected compliance labels to generate an accurate user portrait.

[0095] Through this solution, by obtaining the interaction type, interaction frequency, and interaction depth, and analyzing the user's real-time behavior data, it helps to adjust the service and marketing strategies in a timely manner to meet the rapidly changing market demands. Through matching, determining the centralized associated labels for the user's real-time interaction helps to more accurately locate the user's needs and provide more personalized services. The real-time confidence score helps to evaluate the accuracy and relevance of the centralized associated labels, and at the same time helps to identify which centralized associated labels can better reflect the user's true needs. Through screening, removing the labels that are no longer centrally associated with the user and retaining the centralized associated labels that accurately reflect the user's current state helps to maintain the accuracy and timeliness of the user portrait. The accurate user portrait reflects the user's latest behaviors and preferences, provides a basis for more accurate market positioning and personalized services, helps to improve customer satisfaction and loyalty, and enhances market competitiveness.

[0096] In some embodiments, analyze the behavior data to determine the high-frequency behavior label set; construct a public label set and a potential label set according to the public needs and potential needs; calculate the first set similarity and the second set similarity between the high-frequency behavior label set and the public label set and the potential label set respectively; compare the first set similarity and the second set similarity with the preset similarity threshold respectively. If any one of the first set similarity and the second set similarity is less than the preset similarity threshold, extract the temporal behavior sequence from the behavior data; according to the temporal behavior sequence, eliminate the overlapping behaviors in the public label set and the potential label set; input the eliminated public label set and the eliminated potential label set into a preset demand prediction model to predict the implicit characteristics.

[0097] The high-frequency behavior label set can be a set of the most common behavior data labels of the user identified by analyzing the behavior data.

[0098] The public label set can be a set of labels obtained based on market research and behavioral data analysis, representing the common needs of users.

[0099] The potential label set can be a set of potential demand labels of users inferred by analyzing behavioral data and market trends.

[0100] The first set similarity can be the similarity between the high-frequency behavior label set and the public label set.

[0101] The second set similarity can be the similarity between the high-frequency behavior label set and the potential label set.

[0102] The preset similarity threshold can be a similarity threshold preset for determining whether the similarity between the high-frequency behavior label set and the public label set or the potential label set is high enough. It is stored in the server in advance and called when in use.

[0103] The time-series behavior sequence can be the behavior sequence of a user within a period of time.

[0104] The overlapping behavior can be the same behavior that exists in both the public label set and the potential label set.

[0105] The preset demand prediction model can be a model pre-constructed for predicting users' potential demands. It is stored in the server in advance and called when in use.

[0106] Specifically, use data mining techniques to analyze the behavioral data of users, identify high-frequency behaviors, and create a high-frequency behavior label set. Construct a public label set and a potential label set according to public demands and potential demands. Use a similarity calculation method to calculate the first set similarity between the high-frequency behavior label set and the public label set. Use a similarity calculation method to calculate the second set similarity between the high-frequency behavior label set and the potential label set. Compare the calculated first set similarity and second set similarity with the preset similarity threshold. If either the first set similarity or the second set similarity is less than the preset similarity threshold, extract the time-series behavior sequence from the behavioral data of the user. Use time series analysis techniques to analyze the time-series behavior sequence. According to the analysis results of the time-series behavior sequence, identify the overlapping behaviors that already exist in the public label set and the potential label set. According to the identified overlapping behaviors, use data cleaning techniques to eliminate the overlapping behaviors in the public label set and the potential label set. Input the eliminated public label set and potential label set into the preset demand prediction model and use machine learning algorithms for prediction. According to the prediction results, thus identify the implicit characteristics.

[0107] Through this solution, the high-frequency behavior tag set helps to identify users' daily behavior habits and preferences, thereby better positioning and segmenting user groups. Constructing the mass tag set and potential tag set according to the mass demand and potential demand helps with market segmentation and personalized marketing. By calculating the first set similarity and the second set similarity between the high-frequency behavior tag set and the mass tag set, and the potential tag set, it is determined whether it is necessary to further analyze the user's behavior data. By comparing the first set similarity and the second set similarity with the preset similarity threshold, it is judged whether the user's behavior data is consistent with the preset demand characteristics. If the first set similarity and the second set similarity are lower than the preset similarity threshold, it indicates that the user's behavior data is inconsistent with the preset demand characteristics, and it is necessary to further analyze the user's time-series behavior sequence. Extracting the time-series behavior sequence in the behavior data helps to discover the trends and changes in user behavior, thereby better identifying the user's needs and behavior data. Removing the overlapping behaviors in the mass tag set and the potential tag set helps to simplify the preset demand prediction model and improve the efficiency and accuracy of the preset demand prediction model. Inputting the mass tag set and the potential tag set after removal into the preset demand prediction model to predict the user's implicit characteristics helps to more comprehensively identify the user, formulate personalized service and marketing strategies, and improve user satisfaction.

[0108] In some embodiments, the event type and behavior timestamp in the behavior data are extracted; according to the event type, the FP-growth algorithm is used to mine high-frequency behavior combinations to determine the behavior pattern primitives; the behavior timestamp is analyzed to calculate the time interval distribution of adjacent behaviors; according to the time interval distribution, the behavior time characteristics are determined; according to the behavior pattern primitives and the behavior time characteristics, the unique characteristics are determined.

[0109] The event type can be a specific event of the user on the platform.

[0110] The behavior timestamp can be the specific time point when the user performs some behaviors.

[0111] The FP-growth algorithm can be a frequent item set mining algorithm for efficiently discovering frequent item sets and association rules from a large amount of behavior data.

[0112] The high-frequency behavior combination can be a frequently occurring user behavior combination identified by the FP-growth algorithm.

[0113] The behavior pattern primitive can be the basic unit of the user behavior pattern.

[0114] Adjacent behaviors can be two or more consecutive user behaviors in the time series.

[0115] The time interval distribution can be the distribution of the time intervals between adjacent behaviors.

[0116] The behavioral time characteristics can be time rules and patterns such as the active period, behavioral cycle, response time, etc. of household behaviors.

[0117] Specifically, use data extraction technology to extract event types and behavioral timestamps from behavioral data. Use the FP-growth algorithm to perform frequent itemset mining on event types to identify high-frequency behavior combinations. Determine behavioral pattern primitives based on the high-frequency behavior combinations. Calculate the time difference between adjacent behaviors in the time-series behavior sequence according to the behavioral timestamps. Aggregate the time interval data of the user group to construct a time interval distribution. Analyze the time interval distribution to identify the behavioral time characteristics of users. Combine the behavioral pattern primitives and behavioral time characteristics to determine the unique characteristics of users.

[0118] Through this solution, by extracting event types and behavioral timestamps, identifying users' interaction behaviors and time patterns, it helps to better analyze users' behavioral habits. Behavioral pattern primitives reveal common interaction patterns of users, which helps to identify users' behavioral preferences. Analyze the behavioral timestamps, calculate the time interval distribution between adjacent behaviors, and analyze the time rules of users' behaviors. The time interval distribution reveals the time patterns of users' behaviors, which helps to identify users' behavioral cycles and active time periods. According to the time interval distribution, determine the behavioral time characteristics of users, such as active periods, behavioral cycles, etc., which helps to analyze users' behavioral patterns and time preferences, providing a basis for personalized services. Combine the behavioral pattern primitives and behavioral time characteristics to determine the unique characteristics of users, reflecting users' personalized behavioral patterns, which helps to more deeply identify users' personalized needs and provide support for customized services and marketing strategies.

[0119] In some embodiments, determine the conflict type according to the mutation candidate tags and several implicit tags; analyze the conflict type to determine the conflict pattern; determine the template adjustment scheme according to the conflict pattern and the tag conflict rate; adjust the initial gene template according to the template adjustment scheme.

[0120] The conflict type can be a situation of inconsistency between the mutation candidate tags and several implicit tags during the construction of the user portrait.

[0121] The conflict pattern can be co-occurrence patterns, dependency relationships, competitive relationships, etc. presented by the conflict type during the construction of the user portrait.

[0122] The template adjustment scheme can be an adjustment scheme for the user portrait construction template to identify conflicts in the user portrait.

[0123] Specifically, analyze the mutant candidate labels generated by the genetic algorithm and the implicit labels of the user. Based on the analysis results between the mutant candidate labels and several implicit labels, determine the conflict situation, and thus identify the conflict type. Use statistical analysis techniques to analyze the conflict type, and based on the analysis results, determine the conflict pattern. Analyze the identified conflict pattern, and at the same time analyze the label conflict rate, that is, the frequency and severity of conflicts between the mutant candidate labels and several implicit labels. Based on the analysis results of the conflict pattern and the label conflict rate, formulate a template adjustment strategy. Adjust the initial gene template according to the template adjustment plan.

[0124] Through this solution, by analyzing the conflicts between the mutant candidate labels and several implicit labels, the inconsistencies between the mutant candidate labels and several implicit labels are identified, providing a basis for conflict pattern analysis. By analyzing the conflict type, the conflict pattern is identified, reflecting the mutual relationship between the mutant candidate labels and several implicit labels, providing guidance for the template adjustment plan. According to the conflict pattern and the label conflict rate, formulate a template adjustment plan to improve the accuracy and consistency of the user portrait. Adjust the initial gene template according to the template adjustment plan, thereby optimizing the construction process of the user portrait.

[0125] In some embodiments, continuously monitor the continuous interaction data of the target user; analyze the continuous interaction data to determine the updated data; according to the updated data, dynamically expand the multi-branch candidate label set of the precise user portrait, and perform pruning optimization on the expanded multi-branch candidate label set.

[0126] The continuous interaction data can be the data generated by the user's continuous interaction behavior on the platform.

[0127] The updated data can be the data analyzed from the continuous interaction data and needs to be updated to the user portrait.

[0128] Specifically, use real-time data stream processing technology to continuously monitor the continuous interaction data of the user. Use data mining and analysis techniques to deeply analyze the continuous interaction data to determine the updated data. According to the updated data, dynamically expand the multi-branch candidate label set of the precise user portrait, and use the pruning process in the genetic algorithm to optimize the expanded multi-branch candidate label set.

[0129] Through this solution, by continuously monitoring the continuous interaction data of users and analyzing the user behavior patterns in real time, it helps to adjust service strategies in a timely manner to adapt to the rapidly changing market demands. By analyzing the continuous interaction data of users and identifying the user portrait data that needs to be updated, it helps to adjust the user portrait in a timely manner, reflect the latest behaviors and preferences of users, and improve the accuracy and timeliness of the user portrait. According to the updated data, dynamically expanding the multi-branch candidate tag set of the precise user portrait helps to identify the diverse needs and behaviors of users and provide support for personalized services and marketing strategies. Using the pruning process in the genetic algorithm to optimize the expanded multi-branch candidate tag set and removing redundant or irrelevant candidate tags helps to simplify the user portrait, reduce redundant information, and improve the efficiency and practicality of the user portrait.

[0130] Figure 3 The following is a schematic structural diagram of an enterprise digital customer relationship management system provided by an embodiment of the present application, as Figure 3 shown, the enterprise digital customer relationship management system 300 of this embodiment includes: a data acquisition module 301, a data analysis module 302, an initial portrait generation module 303, and a portrait optimization module 304.

[0131] The data acquisition module 301 is configured to acquire the attribute data and behavior data of the target user; The data analysis module 302 is configured to analyze the attribute data and the behavior data to generate a plurality of independent portrait modules; The initial portrait generation module 303 is configured to combine the independent portrait modules through a genetic algorithm to generate an initial user portrait including a multi-branch candidate tag set; The portrait optimization module 304 is configured to acquire the real-time interaction data of the target user, and perform pruning optimization on the multi-branch candidate tag set in the initial user portrait according to the real-time interaction data to obtain a precise user portrait for customer relationship management.

[0132] Optionally, when the data analysis module 302 analyzes the attribute data and the behavior data to generate a plurality of independent portrait modules, it is configured to: Analyze the attribute data to determine the basic characteristics; Analyze the behavior data to determine the unique characteristics; Predict the implicit characteristics according to the basic characteristics and the unique characteristics; Determine the basic portrait module according to the basic characteristics; Determine the unique portrait module according to the unique characteristics; Determine the implicit portrait module according to the implicit characteristics; Use the basic portrait module, the unique portrait module, and the implicit portrait module as a plurality of independent portrait modules.

[0133] Optionally, the attribute data includes age, gender, and geographical location. When predicting the implicit characteristics based on the basic characteristics and the unique characteristics, the data analysis module 302 is configured to: Analyze the behavior data to determine click behavior, purchase behavior, and dwell time; Analyze the age and gender to determine the general needs; Analyze the geographical location, the click behavior, the purchase behavior, and the dwell time to determine potential needs; Predict the implicit characteristics based on the general needs and the potential needs.

[0134] Optionally, when generating an initial user portrait containing a multi-branch candidate tag set by combining the independent portrait modules through a genetic algorithm, the initial portrait generation module 303 is configured to: Determine an initial gene template according to the basic portrait module; Analyze the unique portrait modules through a genetic algorithm to determine the co-occurrence probability of any two unique portrait modules; Determine the module connection between any two unique portrait modules according to the co-occurrence probability; Randomly cross-combine the unique portrait modules according to the module connection to generate mutant candidate tags; Determine a number of implicit tags according to the implicit portrait module; Analyze the mutant candidate tags and the number of implicit tags to determine the tag conflict rate; Adjust the initial gene template according to the tag conflict rate to generate an initial user portrait containing a multi-branch candidate tag set.

[0135] Optionally, when pruning and optimizing the multi-branch candidate tag set in the initial user portrait according to the real-time interaction data to obtain an accurate user portrait, the portrait optimization module 304 is configured to: Extract the interaction type, interaction frequency, and interaction depth from the real-time interaction data; Match the interaction type with the multi-branch candidate tags to determine the concentrated associated tags; Calculate the real-time confidence score of the concentrated associated tags based on the interaction frequency and interaction depth; Compare the real-time confidence score with a preset dynamic threshold, and screen the concentrated associated tags according to the comparison result to obtain qualified tags; Generate the accurate user portrait according to the qualified tags.

[0136] Optionally, when the data analysis module 302 predicts implicit features based on the public demand and the potential demand, it is used for: Analyze the behavioral data to determine a set of high-frequency behavioral tags; Construct a set of public tags and a set of potential tags according to the public demand and the potential demand; Calculate a first set similarity and a second set similarity between the set of high-frequency behavioral tags and the set of public tags and the set of potential tags respectively; Compare the first set similarity and the second set similarity with a preset similarity threshold respectively. If any one of the first set similarity and the second set similarity is less than the preset similarity threshold, extract the time-series behavioral sequence in the behavioral data; According to the time-series behavioral sequence, eliminate the overlapping behaviors in the set of public tags and the set of potential tags; Input the set of public tags after elimination and the set of potential tags after elimination into a preset demand prediction model to predict implicit features.

[0137] Optionally, when the data analysis module 302 analyzes the behavioral data to determine unique features, it is used for: Extract the event type and behavioral timestamp in the behavioral data; According to the event type, mine high-frequency behavioral combinations through the FP-growth algorithm to determine behavioral pattern primitives; Analyze the behavioral timestamp and calculate the time interval distribution of adjacent behaviors; Determine behavioral time characteristics according to the time interval distribution; Determine unique features according to the behavioral pattern primitives and the behavioral time characteristics.

[0138] Optionally, when the initial portrait generation module 303 adjusts the initial gene template according to the tag conflict rate, it is used for: Determine the conflict type according to the mutant candidate tags and the several implicit tags; Analyze the conflict type to determine the conflict pattern; Determine a template adjustment scheme according to the conflict pattern and the tag conflict rate; Adjust the initial gene template according to the template adjustment scheme.

[0139] Optionally, the enterprise digital customer relationship management system further includes a portrait expansion module 305, which is used for: Continuously monitor the continuous interaction data of the target user; Analyze the continuous interaction data to determine updated data; Dynamically expand the multi-branch candidate tag set of the precise user profile according to the updated data, and perform pruning optimization on the expanded multi-branch candidate tag set.

[0140] The system of this embodiment can be used to execute the method of any of the above embodiments, and its implementation principle and technical effects are similar, which will not be elaborated here.

Claims

1. A method for enterprise digital customer relationship management, characterized in that, Including: Obtain the attribute data and behavior data of the target user; Analyze the attribute data and the behavior data to generate multiple independent portrait modules; Combine the independent portrait modules through a genetic algorithm to generate an initial user portrait containing a multi-branch candidate tag set; Obtain the real-time interaction data of the target user, and according to the real-time interaction data, prune and optimize the multi-branch candidate tag set in the initial user portrait to obtain an accurate user portrait for customer relationship management.

2. The method according to claim 1, wherein The analyzing the attribute data and the behavior data to generate multiple independent portrait modules includes: Analyze the attribute data to determine the basic characteristics; Analyze the behavior data to determine the unique characteristics; Predict the implicit characteristics according to the basic characteristics and the unique characteristics; Determine the basic portrait module according to the basic characteristics; Determine the unique portrait module according to the unique characteristics; Determine the implicit portrait module according to the implicit characteristics; Take the basic portrait module, the unique portrait module, and the implicit portrait module as multiple independent portrait modules.

3. The method according to claim 2, wherein The attribute data includes age, gender, and geographical location; the predicting the implicit characteristics according to the basic characteristics and the unique characteristics includes: Analyze the behavior data to determine the click behavior, purchase behavior, and stay duration; Analyze the age and gender to determine the general needs; Analyze the geographical location, the click behavior, the purchase behavior, and the stay duration to determine the potential needs; Predict the implicit characteristics according to the general needs and the potential needs.

4. The method according to claim 2, wherein The combining the independent portrait modules through a genetic algorithm to generate an initial user portrait containing a multi-branch candidate tag set includes: Determine the initial gene template according to the basic portrait module; Through a genetic algorithm, analyze the unique portrait modules to determine the co-occurrence probability of any two unique portrait modules; Determine the module connection between any two unique portrait modules according to the co-occurrence probability; Randomly cross-combine the unique portrait modules according to the module connection to generate mutant candidate tags; Determine several implicit tags according to the implicit portrait module; Analyze the mutant candidate tags and the several implicit tags to determine the tag conflict rate; Adjust the initial gene template according to the tag conflict rate to generate an initial user portrait containing a multi-branch candidate tag set.

5. The method according to claim 1, characterized in that The pruning and optimizing the multi-branch candidate tag set in the initial user portrait according to the real-time interaction data to obtain an accurate user portrait includes: Extract the interaction type, interaction frequency, and interaction depth from the real-time interaction data; Match the interaction type with the multi-branch candidate tags to determine the concentrated associated tags; Calculate the real-time confidence score of the concentrated associated tags based on the interaction frequency and interaction depth; Compare the real-time confidence score with a preset dynamic threshold, and screen the concentrated associated tags according to the comparison result to obtain qualified tags; Generate the accurate user portrait according to the qualified tags.

6. The method according to claim 3, wherein The predicting the implicit characteristics according to the general needs and the potential needs includes: Analyze the behavior data to determine the high-frequency behavior tag set; Construct a set of popular tags and a set of potential tags according to the said popular demands and potential demands; Calculate the first set similarity and the second set similarity between the set of high-frequency behavior tags and the set of popular tags and the set of potential tags respectively; Compare the first set similarity and the second set similarity with a preset similarity threshold respectively. If any one of the first set similarity and the second set similarity is less than the preset similarity threshold, extract the sequential behavior sequence in the behavior data; According to the sequential behavior sequence, eliminate the overlapping behaviors in the set of popular tags and the set of potential tags; Input the set of popular tags after elimination and the set of potential tags after elimination into a preset demand prediction model to predict the implicit characteristics.

7. The method according to claim 6, wherein The analyzing the behavior data to determine unique characteristics includes: Extract the event type and behavior timestamp in the behavior data; According to the event type, mine high-frequency behavior combinations through the FP-growth algorithm to determine the behavior pattern primitives; Analyze the behavior timestamp and calculate the time interval distribution of adjacent behaviors; According to the time interval distribution, determine the behavior time characteristics; According to the behavior pattern primitives and the behavior time characteristics, determine the unique characteristics.

8. The method according to claim 4, characterized in that The adjusting the initial gene template according to the tag conflict rate includes: Determine the conflict type according to the mutation candidate tags and the several implicit tags; Analyze the conflict type to determine the conflict pattern; According to the conflict pattern and the tag conflict rate, determine the template adjustment scheme; According to the template adjustment scheme, adjust the initial gene template.

9. The method according to claim 1, characterized in that The method further includes: Continuously monitor the continuous interaction data of the target user; Analyze the continuous interaction data to determine the updated data; According to the updated data, dynamically expand the multi-branch candidate tag set of the precise user portrait and perform pruning optimization on the expanded multi-branch candidate tag set.

10. An enterprise digital customer relationship management system, characterized in that, Applied to the method according to any one of claims 1-9, includes: A data acquisition module for acquiring the attribute data and behavior data of the target user; A data analysis module for analyzing the attribute data and the behavior data to generate multiple independent portrait modules; An initial portrait generation module for combining the independent portrait modules through a genetic algorithm to generate an initial user portrait containing a multi-branch candidate tag set; A portrait optimization module for acquiring the real-time interaction data of the target user and performing pruning optimization on the multi-branch candidate tag set in the initial user portrait according to the real-time interaction data to obtain a precise user portrait for customer relationship management.

Images