Active customer screening method and system based on multi-source heterogeneous data

Abstract

The invention is suitable for the technical field of customer screening, and particularly relates to an active customer screening method and system based on multi-source heterogeneous data, and the method comprises the steps: S100: obtaining all indexes in the multi-source heterogeneous data, determining the data source of each index, and calculating the interaction rate of customers in each index according to the data source, branch nodes in one-to-one correspondence with the indexes are created, the branch nodes are mounted to root nodes which are constructed in advance, and each client corresponds to one root node; s200, collecting registration information of a client; according to the method, the active degree of the customer can be visually displayed by constructing the screening tree, meanwhile, the active customer is determined from multiple dimensions according to different weights of customer behaviors, the screening efficiency is greatly improved, the active condition of each customer can be determined by constructing the screening forest, the customer behaviors are more comprehensively evaluated, and the user experience is improved. And the client hierarchical management efficiency is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of customer screening, and in particular to an active customer screening method and system based on multi-source heterogeneous data. Background Art

[0002] Multi-source heterogeneous data refers to data from different data sources, formats and types. By utilizing multi-source heterogeneous data, we can conduct a comprehensive analysis of customer behaviors and characteristics, thereby screening out highly active customers. Taking e-commerce platforms as an example, measuring customer activity only by the number of orders is very one-sided, because it cannot fully reflect the customer's potential value. Customers may recommend used items to friends through the social modules pre-integrated in the e-commerce platform.

[0003] In other words, different business goals need to focus on different customer behaviors, which requires setting different weights for each customer behavior; therefore, "how to dynamically define active customers based on weights" is the technical problem that the present invention needs to solve. Summary of the invention

[0004] The purpose of the present invention is to provide an active customer screening method and system based on multi-source heterogeneous data to solve the problem of "how to dynamically define active customers based on weights" raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: An active customer screening method based on multi-source heterogeneous data, the method comprising: S100: Obtain all indicators in multi-source heterogeneous data, determine the data source of each indicator, and calculate the customer interaction rate in each indicator based on the data source, create branch nodes corresponding to the indicators one by one, and mount the branch nodes to the pre-built root nodes, where each customer corresponds to a root node; S200: Collecting customer registration information, uploading the registration information to a corresponding root node, integrating branch nodes and the root node, generating a screening tree, and inserting a tag generated by a data source into the screening tree; S300: Integrate the screening trees of all customers to generate a screening forest, build a data interface for the screening forest, write a preset comparison table into the screening forest, wherein the comparison table at least includes a weight value of each indicator, calculate the activity value of each screening tree through the weight value and the interaction rate, configure a corresponding relationship between the activity value and the tilt angle, and tilt the screening tree; S400: Obtain screening requirements, traverse a target tree from the screening forest, define the target tree as active customers, integrate the indicators and registration information of all active customers, generate a screening form, and send the screening form to a preset terminal.

[0006] Furthermore, the S100 includes: Collecting real-time data of each indicator via the data source, and preprocessing the real-time data; Construct a time window, count the number of customer interactions within the time window, and calculate the interaction frequency.

[0007] Furthermore, the S100 further includes: Using the real-time data, dynamically update all branch nodes; With time as the horizontal axis and interaction frequency as the vertical axis, draw a trend graph to identify high and low interaction periods.

[0008] Further, the S200 includes: configuring the interrelationships between the indicators and generating a balance factor; The balance factor is embedded into the screening tree, and the weight value in the screening tree is adjusted.

[0009] Further, the S300 includes: Based on the weight value and interaction rate of each indicator, the score of each indicator is calculated and uploaded to the corresponding branch node; The scores of all the branch nodes in the screening tree are superimposed to obtain the activity value of the screening tree.

[0010] Furthermore, the S300 further includes: Configure an active threshold for each indicator, define customers with scores greater than the active threshold as high-frequency customers, and integrate the high-frequency customers into the screening form; Construct a push mechanism based on the score.

[0011] Further, the S400 includes: Integrate each customer's metrics and registration information to generate a multi-dimensional portrait and cluster customers into several categories; Based on the real-time data within the indicator, the multi-dimensional portrait is optimized, the associations between categories are traversed, and an association network is constructed.

[0012] Furthermore, the system comprises: The mounting module is used to obtain all indicators in multi-source heterogeneous data, determine the data source of each indicator, and calculate the customer interaction rate in each indicator based on the data source, create branch nodes corresponding to the indicators one by one, and mount the branch nodes to the pre-built root nodes, where each customer corresponds to a root node; An insertion module, used to collect customer registration information, upload the registration information to a corresponding root node, integrate branch nodes and root nodes, generate a screening tree, and insert tags generated by a data source into the screening tree; A tilt module is used to integrate the screening trees of all customers, generate a screening forest, build a data interface for the screening forest, write a preset comparison table into the screening forest, wherein the comparison table at least includes a weight value of each indicator, calculate the activity value of each screening tree through the weight value and the interaction rate, configure the corresponding relationship between the activity value and the tilt angle, and tilt the screening tree; The sending module is used to obtain the screening requirements, traverse the target tree from the screening forest, define the target tree as active customers, integrate the indicators and registration information of all active customers, generate a screening form, and send the screening form to a preset terminal.

[0013] Furthermore, the mounting module includes: A preprocessing unit, used to collect real-time data of each indicator via the data source, and preprocess the real-time data; A calculation unit, used to construct a time window, count the number of customer interactions within the time window, and calculate the interaction frequency; An updating unit, used to dynamically update all branch nodes using the real-time data; The drawing unit is used to draw a change trend graph with time as the horizontal axis and the interaction frequency as the vertical axis, so as to identify the high interaction period and the low interaction period.

[0014] Furthermore, the insertion module includes: A generating unit, used to configure the mutual relationship between the indicators and generate a balance factor; An adjusting unit is used to embed the balance factor into the screening tree and adjust the weight value in the screening tree.

[0015] Compared with the prior art, the present invention has the following beneficial effects: By determining indicators and data sources, the comprehensiveness and accuracy of the data can be improved, covering more comprehensive customer behaviors and enhancing the representativeness of the data. By determining the interaction rate, active customers can be refined and a data basis can be provided for the screening process. By constructing a screening tree, the customer's activity level can be intuitively displayed. At the same time, according to the different weights of customer behaviors, active customers can be determined from multiple dimensions, which greatly improves the screening efficiency. By constructing a screening forest, the activity status of each customer can be determined, and customer behaviors can be evaluated more comprehensively, greatly improving the efficiency of customer stratification management. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1A flowchart of an active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention; Figure 2 A first sub-flow chart of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention; Figure 3 A second sub-flow chart of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention; Figure 4 A third sub-flow diagram of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention; Figure 5 A fourth sub-flow chart of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention; Figure 6 A block diagram of the composition of an active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention; Figure 7 A block diagram of the components of the mounting module in the active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention; Figure 8 A block diagram of the composition of the insertion module in the active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention; Fig. 9 A block diagram of the composition of a tilt module in an active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention; Fig.10 A block diagram of the composition of a sending module in an active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] In Example 1, Figure 1 The implementation process of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention is shown and described in detail below: S100: Obtain all indicators in multi-source heterogeneous data, determine the data source of each indicator, and calculate the customer interaction rate in each indicator based on the data source, create branch nodes corresponding to the indicators one by one, and mount the branch nodes to the pre-built root nodes, where each customer corresponds to a root node.

[0019] Determine the indicators needed for customer screening, such as order volume, comment volume, and recommendation times, and determine the data source for each indicator, where the data source can be a single platform or multiple platforms, or multiple modules in a single platform; determine the customer's interaction rate over a period of time, which is the frequency of likes, comments, and sharing.

[0020] Create a branch node for each indicator of the customer and synchronize the interaction rate to the branch node. Create a root node for each customer and mount the branch node corresponding to each indicator to the root node.

[0021] For example, Customer A often places orders and comments on a certain platform, but does not perform any other operations. Based on the number of orders and comments placed by Customer A over a period of time, the customer's interaction rate is determined, and two branch nodes are generated. The interaction rate is uploaded to the corresponding branch node, and the branch node is mounted to the root node corresponding to Customer A.

[0022] S200: Collect the customer's registration information, upload the registration information to the corresponding root node, integrate the branch nodes and the root node, generate a filter tree, and insert a tag generated by the data source into the filter tree.

[0023] In the preset platform, the customer's registration information is collected, which should include the customer's age and preferences, etc. The registration information is uploaded to the root node corresponding to the customer, the root node and branch nodes corresponding to each customer are integrated to obtain a filter tree, and labels are inserted; the advantage of inserting labels is that the data source can be quickly traced back and the original data can be determined.

[0024] S300: Integrate the screening trees of all customers to generate a screening forest, build a data interface for the screening forest, write a preset comparison table into the screening forest, wherein the comparison table includes at least a weight value for each indicator, calculate the active value of each screening tree through the weight value and the interaction rate, configure the corresponding relationship between the active value and the tilt angle, and tilt the screening tree.

[0025] Integrate all the screening trees to generate a screening forest, and build a data interface in the screening forest. Through the data interface, write real-time data into the screening forest, determine the weight value of each indicator, and determine the activity value of each screening tree based on the weight value and interaction rate of each indicator.

[0026] Integrate the active values ​​of all customers, determine the fluctuation range of the active values, divide the active values ​​into multiple intervals, determine a tilt angle for each interval, and adjust the customer's screening tree to the corresponding tilt angle.

[0027] S400: Obtain screening requirements, traverse a target tree from the screening forest, define the target tree as active customers, integrate the indicators and registration information of all active customers, generate a screening form, and send the screening form to a preset terminal.

[0028] Determine the screening requirements, where the screening requirements can be the number of customers and active values, etc. Use the screening requirements to select a target tree from the screening forest, write the indicators and registration information in the target tree into the template, generate a screening form, and send the screening form to a preset terminal, where the preset terminal can be a screening organizer terminal.

[0029] In Example 2, Figure 2 The implementation process of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention is shown, and S100 is described in detail below: S101: Collecting real-time data of each indicator via the data source, and preprocessing the real-time data.

[0030] The real-time data of each indicator is read from the data source and preprocessed. The specific steps of preprocessing include: cleaning, format conversion and standardization of real-time data.

[0031] S102: construct a time window, count the number of customer interactions within the time window, and calculate the interaction frequency.

[0032] Determine the initial time and end time, build a time window, and count the number of customer interactions in each indicator within the time window; calculate the interaction frequency based on the difference between the end time and the initial time, combined with the number of interactions, where each indicator corresponds to an interaction frequency.

[0033] In Example 3, Figure 2 The implementation process of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention is shown. S100 is further described in detail below, as follows: S103: Utilize the real-time data to dynamically update all branch nodes.

[0034] Use real-time data to dynamically update the interaction rate in branch nodes.

[0035] S104: With time as the horizontal axis and interaction frequency as the vertical axis, draw a change trend graph to identify high interaction periods and low interaction periods.

[0036] According to the corresponding relationship between time and interaction frequency in the time window, a change trend graph is drawn with time as the horizontal axis and interaction frequency as the vertical axis. The dividing value is determined in the change trend graph, and the part with an interaction rate greater than the dividing value is found in the change trend graph. The time period corresponding to this part is determined as a high interaction period, otherwise, it is determined as a low interaction period.

[0037] In Example 4, Figure 3 The implementation process of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention is shown, and S200 is described in detail below, as follows: S201: Configure the relationship between the indicators and generate a balance factor.

[0038] Configure the relationship between indicators, among which the balance factor is the factor that affects the interaction rate.

[0039] S202: embed the balance factor into the screening tree, and adjust the weight value in the screening tree.

[0040] Determine the influencing factors of the interaction rate in each indicator, and use this influencing factor to adjust the weight value of each indicator. The influencing factors include: customer age, gender, and region. In other words, customers of different age groups must have different likes, reviews, and orders. By adding a balancing factor to each screening tree, bias can be eliminated and the true behavioral tendencies of each customer can be evaluated more fairly.

[0041] In Example 5, Figure 4 The implementation process of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention is shown, and S300 is described in detail below: S301: Based on the weight value and interaction rate of each indicator, the score of each indicator is calculated, and the score is uploaded to the corresponding branch node.

[0042] The weight value of each indicator is multiplied by the interaction rate to calculate the score of each indicator, where the score is used to represent the customer's activity level in this indicator.

[0043] S302: Superimpose the scores of all branch nodes in the screening tree to obtain the activity value of the screening tree.

[0044] The scores of all indicators corresponding to the customer are superimposed to obtain the active value of the screening tree.

[0045] In Example 6, Figure 4 The implementation process of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention is shown. S300 is further described in detail below, as follows: S303: Configure an active threshold for each indicator, define the customers whose scores are greater than the active threshold as high-frequency customers, and integrate the high-frequency customers into the screening form.

[0046] For each indicator, determine the active threshold. If the score corresponding to the indicator is greater than the active threshold, the customer is defined as a high-frequency customer and integrated into the screening form. In real life, high-frequency customers are customers who like to comment or like. By determining high-frequency customers, the core audience can be identified, thereby improving the efficiency of using the screening results.

[0047] S304: Constructing a push mechanism based on the score.

[0048] Build a push mechanism and push information based on the score. The push mechanism is: when the score is greater than the preset value, push promotion information to the corresponding customer.

[0049] In Example 7, Figure 5 The implementation process of the active customer screening method based on multi-source heterogeneous data provided by an embodiment of the present invention is shown, and S400 is described in detail below, as follows: S401: Integrate the indicators and registration information of each customer, generate a multi-dimensional portrait, and cluster the customers into several categories; Based on the customer's indicators and registration information, a multi-dimensional portrait is generated. The multi-dimensional portrait can be a behavioral portrait that includes the interaction rate, or an interest portrait that includes the customer's browsing content and keywords, etc.; based on the interest portrait, the customers are clustered into several categories, such as: fashion products category and daily necessities category, etc., where the categories are used to characterize the customer's potential preferences.

[0050] S402: Based on the real-time data in the indicator, optimize the multi-dimensional portrait, traverse the associations between categories, and build an association network.

[0051] Determine the association between the categories corresponding to the customers, and build an association network based on this association; association is the relationship between indicators, and using this relationship, the weight value corresponding to each indicator is adjusted; for example, the weight value corresponding to each indicator can be adjusted based on the behavioral profile.

[0052] Figure 6 The structure block diagram of the active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention is shown. The active customer screening system based on multi-source heterogeneous data 1 includes: The mounting module 11 is used to obtain all indicators in multi-source heterogeneous data, determine the data source of each indicator, and calculate the customer interaction rate in each indicator based on the data source, create branch nodes corresponding to the indicators one by one, and mount the branch nodes to the pre-built root nodes, where each customer corresponds to a root node; Insertion module 12, used to collect customer registration information, upload the registration information to the corresponding root node, integrate branch nodes and root nodes, generate a screening tree, and insert tags generated by the data source into the screening tree; The tilt module 13 is used to integrate the screening trees of all customers, generate a screening forest, build a data interface of the screening forest, write a preset comparison table into the screening forest, wherein the comparison table at least includes a weight value of each indicator, calculate the activity value of each screening tree through the weight value and the interaction rate, configure the corresponding relationship between the activity value and the tilt angle, and tilt the screening tree; The sending module 14 is used to obtain screening requirements, traverse a target tree from the screening forest, define the target tree as active customers, integrate the indicators and registration information of all active customers, generate a screening form, and send the screening form to a preset terminal.

[0053] Figure 7 The structure diagram of the active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention is shown, and the mounting module 11 includes: A preprocessing unit 111 is used to collect real-time data of each indicator via the data source and preprocess the real-time data; A calculation unit 112 is used to construct a time window, count the number of customer interactions within the time window, and calculate the interaction frequency; An updating unit 113, used to dynamically update all branch nodes using the real-time data; The drawing unit 114 is used to draw a change trend graph with time as the horizontal axis and the interaction frequency as the vertical axis, so as to identify the high interaction period and the low interaction period.

[0054] Figure 8 The structure diagram of the active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention is shown, and the insertion module 12 includes: A generating unit 121, configured to configure the mutual relationship between the indicators and generate a balance factor; The adjusting unit 122 is used to embed the balance factor into the screening tree and adjust the weight value in the screening tree.

[0055] Fig. 9The structure block diagram of the active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention is shown, and the tilting module 13 includes: The uploading unit 131 is used to calculate the score of each indicator according to the weight value and interaction rate of each indicator, and upload the score to the corresponding branch node; An obtaining unit 132 is used to superimpose the scores of all branch nodes in the screening tree to obtain an activity value of the screening tree; An integration unit 133 is used to configure an active threshold value for each indicator, define the customers whose scores are greater than the active threshold value as high-frequency customers, and integrate the high-frequency customers into the screening form; The push unit 134 is used to build a push mechanism based on the score.

[0056] Fig.10 The structure diagram of the active customer screening system based on multi-source heterogeneous data provided by an embodiment of the present invention is shown, and the sending module 14 includes: A clustering unit 141, for integrating the indicators and registration information of each customer, generating a multi-dimensional portrait, and clustering the customers into several categories; The traversal unit 142 is used to optimize the multi-dimensional portrait according to the real-time data in the indicator, traverse the associations between the categories, and construct an association network.

[0057] The mounting module 11 is mainly used to complete step S100, the inserting module 12 is mainly used to complete step S200, the tilting module 13 is mainly used to complete step S300, and the sending module 14 is mainly used to complete step S400; The preprocessing unit 111 is mainly used to complete step S101, the calculating unit 112 is mainly used to complete step S102, the updating unit 113 is mainly used to complete step S103, and the drawing unit 114 is mainly used to complete step S104; The generating unit 121 is mainly used to complete step S201, and the adjusting unit 122 is mainly used to complete step S202; The uploading unit 131 is mainly used to complete step S301, the obtaining unit 132 is mainly used to complete step S302, the integrating unit 133 is mainly used to complete step S303, and the pushing unit 134 is mainly used to complete step S304; The clustering unit 141 is mainly used to complete step S401, and the traversal unit 142 is mainly used to complete step S402.

[0058] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0059] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0060] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An active customer screening method based on multi-source heterogeneous data, characterized in that: The method comprises: S100: Obtain all indicators in multi-source heterogeneous data, determine the data source of each indicator, and calculate the customer interaction rate in each indicator based on the data source, create branch nodes corresponding to the indicators one by one, and mount the branch nodes to the pre-built root nodes, where each customer corresponds to a root node; S200: Collecting customer registration information, uploading the registration information to a corresponding root node, integrating branch nodes and the root node, generating a screening tree, and inserting a tag generated by a data source into the screening tree; S300: Integrate the screening trees of all customers to generate a screening forest, build a data interface for the screening forest, write a preset comparison table into the screening forest, wherein the comparison table at least includes a weight value of each indicator, calculate the activity value of each screening tree through the weight value and the interaction rate, configure a corresponding relationship between the activity value and the tilt angle, and tilt the screening tree; S400: Obtain screening requirements, traverse a target tree from the screening forest, define the target tree as active customers, integrate the indicators and registration information of all active customers, generate a screening form, and send the screening form to a preset terminal.

2. The method for screening active customers based on multi-source heterogeneous data according to claim 1, characterized in that: The S100 includes: Collecting real-time data of each indicator via the data source, and preprocessing the real-time data; Construct a time window, count the number of customer interactions within the time window, and calculate the interaction frequency.

3. The method for screening active customers based on multi-source heterogeneous data according to claim 2, characterized in that: The S100 further includes: Using the real-time data, dynamically update all branch nodes; With time as the horizontal axis and interaction frequency as the vertical axis, draw a trend graph to identify high and low interaction periods.

4. The method for screening active customers based on multi-source heterogeneous data according to claim 1, characterized in that: The S200 includes: configuring the interrelationships between the indicators and generating a balance factor; The balance factor is embedded into the screening tree, and the weight value in the screening tree is adjusted.

5. The method for screening active customers based on multi-source heterogeneous data according to claim 1, characterized in that: The S300 includes: Based on the weight value and interaction rate of each indicator, the score of each indicator is calculated and uploaded to the corresponding branch node; The scores of all the branch nodes in the screening tree are superimposed to obtain the activity value of the screening tree.

6. The method for screening active customers based on multi-source heterogeneous data according to claim 5, characterized in that: The S300 further includes: Configure an active threshold for each indicator, define customers with scores greater than the active threshold as high-frequency customers, and integrate the high-frequency customers into the screening form; Construct a push mechanism based on the score.

7. The method for screening active customers based on multi-source heterogeneous data according to claim 2, characterized in that: The S400 includes: Integrate each customer's metrics and registration information to generate a multi-dimensional portrait and cluster customers into several categories; Based on the real-time data within the indicator, the multi-dimensional portrait is optimized, the associations between categories are traversed, and an association network is constructed.

8. An active customer screening system based on multi-source heterogeneous data, characterized in that: The system comprises: The mounting module is used to obtain all indicators in multi-source heterogeneous data, determine the data source of each indicator, and calculate the customer interaction rate in each indicator based on the data source, create branch nodes corresponding to the indicators one by one, and mount the branch nodes to the pre-built root nodes, where each customer corresponds to a root node; An insertion module, used to collect customer registration information, upload the registration information to a corresponding root node, integrate branch nodes and root nodes, generate a screening tree, and insert tags generated by a data source into the screening tree; A tilt module is used to integrate the screening trees of all customers, generate a screening forest, build a data interface for the screening forest, write a preset comparison table into the screening forest, wherein the comparison table at least includes a weight value of each indicator, calculate the activity value of each screening tree through the weight value and the interaction rate, configure the corresponding relationship between the activity value and the tilt angle, and tilt the screening tree; The sending module is used to obtain the screening requirements, traverse the target tree from the screening forest, define the target tree as active customers, integrate the indicators and registration information of all active customers, generate a screening form, and send the screening form to a preset terminal.

9. The active customer screening system based on multi-source heterogeneous data according to claim 8, characterized in that: The mounting module includes: A preprocessing unit, used to collect real-time data of each indicator via the data source, and preprocess the real-time data; A calculation unit, used to construct a time window, count the number of customer interactions within the time window, and calculate the interaction frequency; An updating unit, used to dynamically update all branch nodes using the real-time data; The drawing unit is used to draw a change trend graph with time as the horizontal axis and the interaction frequency as the vertical axis, so as to identify the high interaction period and the low interaction period.

10. The active customer screening system based on multi-source heterogeneous data according to claim 8, characterized in that: The insertion module comprises: A generating unit, used to configure the mutual relationship between the indicators and generate a balance factor; An adjusting unit is used to embed the balance factor into the screening tree and adjust the weight value in the screening tree.

Images