Visually configured dynamic customer grouping method realized based on Drools rule engine

Through the Drools rule engine and decision tree model, multi-dimensional personalized grouping of dynamic customers is realized, the grouping lag caused by static information is solved, and the accuracy and efficiency of customer management are improved.

CN118940159APending Publication Date: 2024-11-12BEIJING WISDOM TOOTH TECH CONSULTING CO LTD

Patent Information

Application Number
CN202410996588.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-24
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing customer grouping method has static customer information and cannot be updated as customer information changes, resulting in a single grouping information and lagging up, which cannot meet the personalized service needs of dynamic customers.

Method used

The visual configuration method based on the Drools rule engine is adopted to obtain multi-dimensional attribute information of dynamic customers, formulate and update grouping rules, and use the decision tree model to group dynamic customers to ensure grouping accuracy and timeliness.

Benefits of technology

It realizes high-accuracy grouping of dynamic customers, can update in real time according to changes in multi-dimensional attributes, provides more personalized services, avoids the problems of single information and lagging updates, and improves customer management efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940159A_ABST
    Figure CN118940159A_ABST
Patent Text Reader

Abstract

The invention discloses a dynamic customer grouping method based on visual configuration realized by a Drools rule engine, and belongs to the field of dynamic customer grouping. The invention relates to a dynamic customer grouping method based on visual configuration realized by a Drools rule engine. The method comprises the following steps of: 1, formulating a dynamic customer grouping rule; step 2, establishing a Drools rule engine system; 3, obtaining dynamic customer multi-dimensional attribute information; step 4, determining groups of dynamic clients; and step 5, visually displaying the grouping process. According to the method, the problems that the current grouping method cannot be updated along with the change of the current customer information, and the information is single and the update lags behind when the customers are grouped are solved, and the dynamic customers can be updated along with the change of the multi-dimensional attribute information of the dynamic customers when the dynamic customers are grouped; the situations that dynamic customer grouping information is single and updating lags behind are avoided, and the accuracy of dynamic customer grouping is guaranteed.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic customer grouping, and in particular to a dynamic customer grouping method with visual configuration implemented based on a Drools rule engine. Background Art

[0002] When managing customers, all customers are usually recorded in the same form, and the salesperson marks different customers differently based on the communication with the customers and the customers' account opening and transaction records.

[0003] A Chinese patent with publication number CN105160553A discloses a customer grouping method and device, including: obtaining first customer information of a customer, the first customer information including at least one of communication record information, rating information, account status information and current grouping information, or the first customer information including operation information on the customer; determining the customer's grouping according to the first customer information; moving the customer to the determined grouping, by obtaining the customer's first customer information, determining the customer's grouping according to the first customer information, and moving the customer to the determined grouping, and moving the customer to different groups according to the customer's communication record information, rating information, account status information, current grouping information or operation information on the customer, so as to facilitate sales staff to explore new customers or provide targeted services to old customers, thereby achieving the effect of improving customer management efficiency.

[0004] In actual use, the grouping method of the above patent obtains static customer information, which will not be updated as the current customer information changes. The customer grouping has single information and delayed updates. Therefore, it does not meet the existing needs. In this regard, we propose a dynamic customer grouping method based on visual configuration implemented by the Drools rule engine. Summary of the invention

[0005] The purpose of the present invention is to provide a dynamic customer grouping method with visual configuration based on the Drools rule engine, so that the grouping of dynamic customers is more accurate, so as to realize the ability to quickly provide more personalized services to dynamic customers in multiple dimensions, and to obtain multi-dimensional attribute information of dynamic customers by setting time intervals. It is possible to regroup according to changes in the multi-dimensional attribute information of dynamic customers, so that when dynamic customers are grouped, they can be updated as the multi-dimensional attribute information of dynamic customers changes, avoiding the problem of solving the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a dynamic customer grouping method based on visual configuration implemented by Drools rule engine, comprising the following steps:

[0007] Step 1: Obtain dynamic customer grouping requirements and formulate dynamic customer grouping rules based on dynamic customer grouping requirements;

[0008] Step 2: Establish a Drools rule engine system, convert dynamic customer grouping rules into Drools rule files, and store the Drools rule files;

[0009] Step 3: Acquire multi-dimensional attribute information of dynamic customers to be grouped, receive a dynamic customer grouping application sent by a client, and determine the dynamic customer multi-dimensional attribute information corresponding to the dynamic customer grouping application;

[0010] Step 4: Determine the grouping of dynamic customers according to the multi-dimensional attribute information of dynamic customers, and move the dynamic customers to the corresponding grouping;

[0011] Step 5: Receive dynamic customer grouping information and visualize the dynamic customer grouping process.

[0012] Preferably, converting the dynamic customer grouping rules into a Drools rule file and storing the Drools rule file specifically includes:

[0013] The Drools rule engine system is used to obtain multi-dimensional attribute information of dynamic customers, and the abstraction and refinement of the multi-dimensional attribute information are configured into rule elements, which are then configured into specific rules to form a dynamic customer grouping rule library;

[0014] Through the rule language algorithm, the rule content read from the dynamic customer grouping rule library is converted into the Drools language and written into the rule file;

[0015] The dynamic customer grouping rules are deployed to the rule container through the rule engine and the Drools rule files are stored.

[0016] Preferably, the obtaining of multi-dimensional attribute information of dynamic customers specifically includes:

[0017] Acquire dynamic customer multi-dimensional attribute information, perform customer feature recognition on the dynamic customer multi-dimensional attribute information according to a preset customer feature recognition model, and obtain dynamic customer features to be grouped, wherein the preset customer feature recognition model is obtained by training a machine learning model based on customer feature recognition sample data.

[0018] The stacked autoencoder is used to process the dynamic customer features to be grouped, and the target feature vector of each dynamic customer is obtained;

[0019] The target feature vector of each dynamic customer is used to build customer portrait model and obtain multi-dimensional attribute information of dynamic customers.

[0020] Preferably, the multi-dimensional attribute information of the dynamic customer includes at least one of communication record information, rating information, account status information and current grouping information.

[0021] Preferably, the Drools rule engine system specifically includes:

[0022] A rule file generation module is used to dynamically generate a program Drl rule file from the tag rules of dynamic customers using the Drools rule engine;

[0023] The rule calling module is used to load the tag rules of dynamic customers into the tag rule database of dynamic customers of the rule engine. The rule engine finds the corresponding rules through the matcher and sends them to the executor.

[0024] The data preservation module is used to receive the tag rules of dynamic customers, parse the tag rules of dynamic customers, generate rule content that can be recognized by the Drools rule engine, and save it.

[0025] Preferably, the step of obtaining multi-dimensional attribute information of dynamic customers to be grouped specifically includes:

[0026] Obtain customer information and determine customer attributes based on the customer information. Customer attributes are static customers and dynamic customers.

[0027] The customer information with the customer attribute of static customers is removed, and the customer information with the customer attribute of dynamic customers is saved.

[0028] Preferably, determining the grouping of dynamic customers according to the multi-dimensional attribute information of dynamic customers specifically includes:

[0029] According to the multi-dimensional attribute information of dynamic customers, the multi-dimensional attribute information of dynamic customers is obtained and updated in real time, and the dynamic customer categories corresponding to different grouping methods of each dynamic customer are obtained;

[0030] The decision tree grouping model is trained based on the multi-dimensional attribute information of dynamic customers and the dynamic customer categories corresponding to different grouping methods of each dynamic customer combined with the Drools rule file;

[0031] Obtaining a trained decision tree grouping model, wherein the trained decision tree grouping model includes relevant attribute information of different grouping methods;

[0032] Determine the grouping method of dynamic customers to be grouped;

[0033] According to the dynamic customer grouping method, the relevant attribute information of the dynamic customers to be grouped is collected and input into the trained decision tree grouping model to obtain the customer grouping results corresponding to the dynamic customer grouping method.

[0034] Preferably, obtaining the customer grouping result corresponding to the dynamic customer grouping method specifically includes:

[0035] The dynamic customers to be grouped whose communication record information, rating information and account status information are all empty and whose current grouping information indicates that the customer does not belong to any group are set as new customer groups;

[0036] If the communication record information, rating information and account status information of the dynamic customers to be grouped are all empty, and the current grouping information indicates that the customer belongs to the new customer group, and the duration of the dynamic customer in the new customer group is longer than the predetermined duration, it is set as a new resource group;

[0037] When the communication record information or rating information of the dynamic customer to be grouped is not empty and the account status information is empty, the grouping of the dynamic customer is determined according to the communication record information or rating information and the situation is set as one of the intended customer grouping, the trash box grouping and the blacklist grouping;

[0038] The dynamic customers to be grouped whose account status information is not empty and whose account status information indicates that the dynamic customer account registered in the exchange system is not activated are grouped as opened but inactivated customers;

[0039] The dynamic customers to be grouped whose account status information is not empty and whose account status information indicates that the customer's account registered in the exchange system has been activated are set as the customer group with opened and activated accounts.

[0040] Preferably, the step of obtaining multi-dimensional attribute information of dynamic customers and updating it in real time specifically includes:

[0041] Setting a time interval for acquiring multidimensional attribute information of dynamic customers, and acquiring multidimensional attribute information of multiple dynamic customers according to the set time interval for acquiring multidimensional attribute information of dynamic customers;

[0042] Conditionally matching the obtained dynamic customer multi-dimensional attribute information with the dynamic customer grouping rules, and if the dynamic customer grouping rules are met, determining the grouping of the dynamic customer according to the dynamic customer multi-dimensional attribute information;

[0043] If the dynamic customer grouping rules are not met, the customers who do not meet the dynamic customer grouping rules will be eliminated, and the dynamic customers who meet the dynamic customer grouping rules will be updated according to the set dynamic customer multi-dimensional attribute information acquisition time interval.

[0044] Preferably, the training decision tree grouping model specifically includes:

[0045] Combine the multi-dimensional attribute information and the dynamic customer categories corresponding to the different grouping methods of each dynamic customer with the Drools rule file as training data;

[0046] Each non-leaf node represents a test of an attribute information, each branch represents the output of an attribute information in a value range, and each leaf node represents a dynamic customer category corresponding to the dynamic customer grouping method;

[0047] According to the training set data, starting from the root node, child nodes are recursively generated from top to bottom until the leaf node is reached to obtain the trained decision tree grouping model.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The present invention trains a decision tree classification model by utilizing the dynamic customer categories corresponding to different grouping methods of each dynamic customer in combination with a Drools rule file, so that the trained decision tree grouping model has higher accuracy in grouping dynamic customers, thereby enabling the dynamic customers to be quickly provided with more personalized services in multiple dimensions. By setting a time interval to obtain the multidimensional attribute information of the dynamic customers, the dynamic customers can be regrouped according to changes in the multidimensional attribute information of the dynamic customers, so that when the dynamic customers are grouped, they can be updated as the multidimensional attribute information of the dynamic customers changes, thereby avoiding the situation where the dynamic customer grouping information is single and the update is delayed, thereby ensuring the accuracy of the dynamic customer grouping. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A schematic diagram of a dynamic customer grouping method based on a visually configured Drools rule engine according to the present invention;

[0051] Figure 2 It is a schematic diagram of a module of the present invention;

[0052] Figure 3 This is a schematic diagram of the dynamic customer multi-dimensional attribute information update of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] In order to solve the problem that the existing grouping method has a single information and delayed update in actual use, since the customer information obtained is static and will not be updated as the current customer information changes, please refer to Figure 1-Figure 3 , this embodiment provides the following technical solutions:

[0055] A dynamic customer grouping method based on visual configuration implemented by Drools rule engine, comprising the following steps:

[0056] Step 1: Obtain dynamic customer grouping requirements, formulate dynamic customer grouping rules according to dynamic customer grouping requirements, and by formulating grouping rules, establish a Drools rule engine system according to the dynamic customer grouping rules, so as to quickly realize the grouping of dynamic customers;

[0057] Step 2: Establish a Drools rule engine system, obtain the multi-dimensional attribute information of dynamic customers through the Drools rule engine system, and configure the abstraction and refinement of the multi-dimensional attribute information into rule elements, use the rule elements to configure specific rules, and form a dynamic customer grouping rule library. The rule content read from the dynamic customer grouping rule library is converted into Drools language and written into the rule file for storage.

[0058] Step 3: Obtain multi-dimensional attribute information of dynamic customers to be grouped, receive the dynamic customer grouping application sent by the client, and determine the dynamic customer multi-dimensional attribute information corresponding to the dynamic customer grouping application. Group the dynamic customers according to the dynamic customer multi-dimensional attribute information, which can not only ensure the accuracy of the grouping, but also quickly find the corresponding dynamic customers according to the grouping of the dynamic customers, which is convenient for the maintenance of the dynamic customers;

[0059] Step 4: Obtain the multidimensional attribute information of dynamic customers according to their multidimensional attribute information and update it in real time, obtain the dynamic customer categories corresponding to different grouping methods of each dynamic customer, and train the decision tree grouping model in combination with the Drools rule file, collect the attribute information of the dynamic customers to be grouped according to the dynamic customer grouping method, and input it into the trained decision tree grouping model, obtain the customer grouping results corresponding to the dynamic customer grouping method and move the dynamic customers to the corresponding group;

[0060] Step 5: Receive dynamic customer grouping information and visualize the dynamic customer grouping process. By visually displaying the dynamic customer grouping process, the dynamic customer grouping information can be understood.

[0061] Convert dynamic customer grouping rules into Drools rule files and store the Drools rule files, including:

[0062] The Drools rule engine system is used to obtain multi-dimensional attribute information of dynamic customers, and the abstraction and refinement of the multi-dimensional attribute information are configured into rule elements, which are then configured into specific rules to form a dynamic customer grouping rule library;

[0063] Through the rule language algorithm, the rule content read from the dynamic customer grouping rule library is converted into the Drools language and written into the rule file;

[0064] The dynamic customer grouping rules are deployed to the rule container through the rule engine and the Drools rule files are stored.

[0065] Get multi-dimensional attribute information of dynamic customers, including:

[0066] Acquire dynamic customer multi-dimensional attribute information, perform customer feature recognition on the dynamic customer multi-dimensional attribute information according to a preset customer feature recognition model, and obtain dynamic customer features to be grouped, wherein the preset customer feature recognition model is obtained by training a machine learning model based on customer feature recognition sample data.

[0067] The stacked autoencoder is used to process the dynamic customer features to be grouped, and the target feature vector of each dynamic customer is obtained;

[0068] The target feature vector of each dynamic customer is used to build customer portraits to obtain multi-dimensional attribute information of dynamic customers. The target feature data of dynamic customers are obtained and feature processed by stacked autoencoders to obtain target feature vectors of dynamic customers, thereby achieving feature learning through stacked autoencoders. Since the feature data are processed by stacked autoencoders and spectral clustering algorithms, a large number of features can be selected and are no longer limited to fewer features, so that the scope of consideration is wider, which effectively improves the accuracy of dynamic customer grouping.

[0069] The multi-dimensional attribute information of a dynamic customer includes at least one of communication record information, rating information, account status information and current grouping information.

[0070] Drools rule engine system, including:

[0071] A rule file generation module is used to dynamically generate a program Drl rule file from the tag rules of dynamic customers using the Drools rule engine;

[0072] The rule calling module is used to load the tag rules of dynamic customers into the tag rule database of dynamic customers of the rule engine. The rule engine finds the corresponding rules through the matcher and sends them to the executor. By establishing a Drools rule engine system and combining Drools rules to group dynamic customers, the rules for grouping dynamic customers are more comprehensive and flexible, thereby achieving the effect of improving the intelligence of dynamic customer grouping and improving the efficiency of dynamic customer grouping;

[0073] The data preservation module is used to receive the tag rules of dynamic customers, parse the tag rules of dynamic customers, generate rule content that can be recognized by the Drools rule engine, and save it.

[0074] Obtain multi-dimensional attribute information of dynamic customers to be grouped, including:

[0075] Obtain customer information and determine customer attributes based on the customer information. Customer attributes are static customers and dynamic customers.

[0076] The customer information with the customer attribute of static customers is removed, and the customer information with the customer attribute of dynamic customers is saved.

[0077] Determine the grouping of dynamic customers according to the multi-dimensional attribute information of dynamic customers, specifically including:

[0078] According to the multi-dimensional attribute information of dynamic customers, the multi-dimensional attribute information of dynamic customers is obtained and updated in real time. The grouping can be adjusted according to the multi-dimensional attribute information of dynamic customers in different time periods, which ensures the accuracy of dynamic customer grouping and avoids grouping dynamic customers by static information. When the multi-dimensional attribute information of dynamic customers changes, it cannot be corrected and re-grouped in time, resulting in the lag of dynamic customer grouping and inconvenience to dynamic customer management. The dynamic customer category corresponding to the different grouping methods of each dynamic customer is obtained;

[0079] The decision tree grouping model is trained based on the multi-dimensional attribute information of dynamic customers and the dynamic customer categories corresponding to different grouping methods of each dynamic customer combined with the Drools rule file;

[0080] Obtaining a trained decision tree grouping model, wherein the trained decision tree grouping model includes relevant attribute information of different grouping methods;

[0081] Determine the grouping method of dynamic customers to be grouped;

[0082] According to the dynamic customer grouping method, the relevant attribute information of the dynamic customers to be grouped is collected and input into the trained decision tree grouping model to obtain the customer grouping results corresponding to the dynamic customer grouping method, which fully considers the attribute information of multiple dimensions of dynamic customers. At the same time, during model training, the dynamic customer categories corresponding to the different grouping methods of each dynamic customer are combined with the Drools rule file to train the decision tree classification model, so that the trained decision tree grouping model has higher accuracy in grouping dynamic customers, thereby realizing the ability to quickly provide more personalized services to dynamic customers in multiple dimensions.

[0083] Get the customer grouping results corresponding to the dynamic customer grouping method, including:

[0084] The dynamic customers to be grouped whose communication record information, rating information and account status information are all empty and whose current grouping information indicates that the customer does not belong to any group are set as new customer groups;

[0085] If the communication record information, rating information and account status information of the dynamic customers to be grouped are all empty, and the current grouping information indicates that the customer belongs to the new customer group, and the duration of the dynamic customer in the new customer group is longer than the predetermined duration, it is set as a new resource group;

[0086] When the communication record information or rating information of the dynamic customer to be grouped is not empty and the account status information is empty, the grouping of the dynamic customer is determined according to the communication record information or rating information and the situation is set as one of the intended customer grouping, the trash box grouping and the blacklist grouping;

[0087] The dynamic customers to be grouped whose account status information is not empty and whose account status information indicates that the dynamic customer account registered in the exchange system is not activated are grouped as opened but inactivated customers;

[0088] The dynamic customers to be grouped whose account status information is not empty and whose account status information indicates that the customer's account registered in the exchange system has been activated are set as the customer group with opened and activated accounts.

[0089] Obtain multi-dimensional attribute information of dynamic customers and update it in real time, including:

[0090] Setting a time interval for acquiring multidimensional attribute information of dynamic customers, and acquiring multidimensional attribute information of multiple dynamic customers according to the set time interval for acquiring multidimensional attribute information of dynamic customers;

[0091] Conditionally matching the obtained dynamic customer multi-dimensional attribute information with the dynamic customer grouping rules, and if the dynamic customer grouping rules are met, determining the grouping of the dynamic customer according to the dynamic customer multi-dimensional attribute information;

[0092] If the dynamic customer grouping rules are not met, the customers who do not meet the dynamic customer grouping rules will be eliminated, and the dynamic customers who meet the dynamic customer grouping rules will be updated according to the set dynamic customer multi-dimensional attribute information acquisition time interval. By setting the time interval to acquire the multi-dimensional attribute information of the dynamic customers, it is possible to regroup according to the changes in the multi-dimensional attribute information of the dynamic customers, so that when the dynamic customers are grouped, they can be updated as the multi-dimensional attribute information of the dynamic customers changes, avoiding the situation where the dynamic customer grouping information is single and the update is delayed, thereby ensuring the accuracy of the dynamic customer grouping.

[0093] Training a decision tree grouping model, including:

[0094] Combine the multi-dimensional attribute information and the dynamic customer categories corresponding to the different grouping methods of each dynamic customer with the Drools rule file as training data;

[0095] Each non-leaf node represents a test of an attribute information, each branch represents the output of an attribute information in a value range, and each leaf node represents a dynamic customer category corresponding to the dynamic customer grouping method;

[0096] According to the training set data, starting from the root node, child nodes are recursively generated from top to bottom until the leaf node is reached to obtain the trained decision tree grouping model.

[0097] In summary, a dynamic customer grouping method based on visual configuration implemented by the Drools rule engine of the present invention obtains target feature data of dynamic customers, uses a stacked autoencoder to perform feature processing on the target feature data of dynamic customers, and obtains a target feature vector of the dynamic customer, thereby realizing feature learning through the stacked autoencoder. Since the feature data is processed by the stacked autoencoder and the spectral clustering algorithm, a large number of features can be selected, and it is no longer limited by fewer features, so that the scope of consideration is wider, and the accuracy of dynamic customer grouping is effectively provided. By using the dynamic customer category corresponding to the different grouping methods of each dynamic customer combined with the Drools rule file to train the decision tree classification model, the trained decision tree grouping model has a higher accuracy in grouping dynamic customers, thereby realizing that more personalized services can be quickly provided to dynamic customers in multiple dimensions. By setting a time interval to obtain the multidimensional attribute information of the dynamic customer, it can be regrouped according to the change of the multidimensional attribute information of the dynamic customer, so that when the dynamic customer is grouped, it can be updated as the multidimensional attribute information of the dynamic customer changes, avoiding the situation where the dynamic customer grouping information is single and the update is delayed, and ensuring the accuracy of the dynamic customer grouping.

[0098] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0099] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A dynamic customer grouping method based on visual configuration implemented by Drools rule engine, characterized in that: The following steps are involved: Step 1: Obtain dynamic customer grouping requirements and formulate dynamic customer grouping rules based on dynamic customer grouping requirements; Step 2: Establish a Drools rule engine system, convert dynamic customer grouping rules into Drools rule files, and store the Drools rule files; Step 3: Acquire multi-dimensional attribute information of dynamic customers to be grouped, receive a dynamic customer grouping application sent by a client, and determine the dynamic customer multi-dimensional attribute information corresponding to the dynamic customer grouping application; Step 4: Determine the grouping of dynamic customers according to the multi-dimensional attribute information of dynamic customers, and move the dynamic customers to the corresponding grouping; Step 5: Receive dynamic customer grouping information and visualize the dynamic customer grouping process.

2. According to claim 1, a method for dynamic customer grouping based on visual configuration implemented by Drools rule engine, characterized in that: The step of converting the dynamic customer grouping rules into a Drools rule file and storing the Drools rule file specifically includes: The Drools rule engine system is used to obtain multi-dimensional attribute information of dynamic customers, and the abstraction and refinement of the multi-dimensional attribute information are configured into rule elements, which are then configured into specific rules to form a dynamic customer grouping rule library; Through the rule language algorithm, the rule content read from the dynamic customer grouping rule library is converted into the Drools language and written into the rule file; The dynamic customer grouping rules are deployed to the rule container through the rule engine and the Drools rule files are stored.

3. According to claim 2, a method for dynamic customer grouping based on visual configuration implemented by Drools rule engine, characterized in that: The obtaining of multi-dimensional attribute information of dynamic customers specifically includes: Acquire dynamic customer multi-dimensional attribute information, perform customer feature recognition on the dynamic customer multi-dimensional attribute information according to a preset customer feature recognition model, and obtain dynamic customer features to be grouped, wherein the preset customer feature recognition model is obtained by training a machine learning model based on customer feature recognition sample data. The stacked autoencoder is used to process the dynamic customer features to be grouped, and the target feature vector of each dynamic customer is obtained; The target feature vector of each dynamic customer is used to build customer portrait model and obtain multi-dimensional attribute information of dynamic customers.

4. The method for dynamic customer grouping based on visual configuration implemented by Drools rule engine according to claim 3 is characterized in that: The multi-dimensional attribute information of the dynamic customer includes at least one of communication record information, rating information, account status information and current grouping information.

5. The method for dynamic customer grouping based on visual configuration implemented by Drools rule engine according to claim 2, characterized in that: The Drools rule engine system specifically includes: A rule file generation module is used to dynamically generate a program Drl rule file from the tag rules of dynamic customers using the Drools rule engine; The rule calling module is used to load the tag rules of dynamic customers into the tag rule database of dynamic customers of the rule engine. The rule engine finds the corresponding rules through the matcher and sends them to the executor. The data preservation module is used to receive the tag rules of dynamic customers, parse the tag rules of dynamic customers, generate rule content that can be recognized by the Drools rule engine, and save it.

6. The method for dynamic customer grouping based on visual configuration implemented by Drools rule engine according to claim 1, characterized in that: The step of obtaining multi-dimensional attribute information of dynamic customers to be grouped specifically includes: Obtain customer information and determine customer attributes based on the customer information. Customer attributes are static customers and dynamic customers. The customer information with the customer attribute of static customers is removed, and the customer information with the customer attribute of dynamic customers is saved.

7. The method for dynamic customer grouping based on visual configuration implemented by Drools rule engine according to claim 1, characterized in that: The determining of the grouping of dynamic customers according to the multi-dimensional attribute information of dynamic customers specifically includes: According to the multi-dimensional attribute information of dynamic customers, the multi-dimensional attribute information of dynamic customers is obtained and updated in real time, and the dynamic customer categories corresponding to different grouping methods of each dynamic customer are obtained; The decision tree grouping model is trained based on the multi-dimensional attribute information of dynamic customers and the dynamic customer categories corresponding to different grouping methods of each dynamic customer combined with the Drools rule file; Obtaining a trained decision tree grouping model, wherein the trained decision tree grouping model includes relevant attribute information of different grouping methods; Determine the grouping method of dynamic customers to be grouped; According to the dynamic customer grouping method, the relevant attribute information of the dynamic customers to be grouped is collected and input into the trained decision tree grouping model to obtain the customer grouping results corresponding to the dynamic customer grouping method.

8. The method for dynamic customer grouping based on visual configuration implemented by Drools rule engine according to claim 7, characterized in that: The obtaining of the customer grouping result corresponding to the dynamic customer grouping method specifically includes: The dynamic customers to be grouped whose communication record information, rating information and account status information are all empty and whose current grouping information indicates that the customer does not belong to any group are set as new customer groups; If the communication record information, rating information and account status information of the dynamic customers to be grouped are all empty, and the current grouping information indicates that the customer belongs to the new customer group, and the duration of the dynamic customer in the new customer group is longer than the predetermined duration, it is set as a new resource group; When the communication record information or rating information of the dynamic customer to be grouped is not empty and the account status information is empty, the grouping of the dynamic customer is determined according to the communication record information or rating information and the situation is set as one of the intended customer grouping, the trash box grouping and the blacklist grouping; The dynamic customers to be grouped whose account status information is not empty and whose account status information indicates that the dynamic customer account registered in the exchange system is not activated are grouped as opened but inactivated customers; The dynamic customers to be grouped whose account status information is not empty and whose account status information indicates that the customer's account registered in the exchange system has been activated are set as the customer group with opened and activated accounts.

9. The method for dynamic customer grouping based on visual configuration implemented by Drools rule engine according to claim 7, characterized in that: The obtaining of multi-dimensional attribute information of dynamic customers and updating in real time specifically includes: Setting a time interval for acquiring multidimensional attribute information of dynamic customers, and acquiring multidimensional attribute information of multiple dynamic customers according to the set time interval for acquiring multidimensional attribute information of dynamic customers; Conditionally matching the obtained dynamic customer multi-dimensional attribute information with the dynamic customer grouping rules, and if the dynamic customer grouping rules are met, determining the grouping of the dynamic customer according to the dynamic customer multi-dimensional attribute information; If the dynamic customer grouping rules are not met, the customers who do not meet the dynamic customer grouping rules will be eliminated, and the dynamic customers who meet the dynamic customer grouping rules will be updated according to the set dynamic customer multi-dimensional attribute information acquisition time interval.

10. The method for dynamic customer grouping based on visual configuration implemented by Drools rule engine according to claim 7, characterized in that: The training decision tree grouping model specifically includes: Combine the multi-dimensional attribute information and the dynamic customer categories corresponding to the different grouping methods of each dynamic customer with the Drools rule file as training data; Each non-leaf node represents a test of an attribute information, each branch represents the output of an attribute information in a value range, and each leaf node represents a dynamic customer category corresponding to the dynamic customer grouping method; According to the training set data, starting from the root node, child nodes are recursively generated from top to bottom until the leaf node is reached to obtain the trained decision tree grouping model.

Citation Information

Patent Citations

  • Method and device for determining business processing rules based on Drools rule engine

    CN114841797A

  • Customer classification method and device

    CN115510994A

  • Method for dynamically setting user labels in hospital based on Drools rule engine

    CN116700721A

  • Settlement card having locked-in card specific merchant and rule-based authorization for each transaction

    US20240169355A1

Cited By

  • Mobile phone information privacy protection method and system based on big data analysis

    CN119577833A

  • Hybrid grouping method and device, electronic equipment and medium

    CN120596985A

  • Resource processing method and device based on visual resource platform, equipment and medium

    CN121880625A