Customer behavior analysis system for marriage and love platform CRM

Abstract

The present application relates to big data resource service technical field, specifically to customer behavior analysis system for marriage platform CRM, the present application firstly in the first recommendation cycle, based on fuzzy clustering algorithm and portrait information will customer clustering, obtain all recommended customers of target customer, further in every recommendation cycle, according to the interactive information of target customer and every recommended customer, update the membership degree of target customer in every cluster and recommended customer, finally in the current time, according to the time sequence change situation of the membership degree of target customer in every cluster, obtain the purpose clear index of target customer, and then adjust the marriage purpose of target customer, the present application analyzes the interactive behavior change of target customer and the recommended customer of platform in different recommendation cycle, analyzes the potential expectation of target customer, so as to continuously adjust the recommended candidate, and then gradually clear the marriage purpose of target customer, so as to improve the matching effect of marriage platform.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data resource service, in particular to a customer behavior analysis system for a marriage platform CRM. BACKGROUND

[0002] Customer Relationship Management (CRM) is a strategy and technology that helps enterprises manage relationships with customers. In a marriage platform, the CRM system collects and analyzes customer basic portraits, behavior information and interaction data to help the marriage platform understand customer needs and provide personalized and refined matching or recommendation services, thereby improving the success rate of marriage matching.

[0003] However, the user portrait in the traditional marriage CRM system is static and fixed, and cannot adapt to the dynamic evolution characteristics of marriage demands. That is, when the marriage purpose filled by the customer at the initial registration changes over time without being timely changed, the marriage platform may continue to recommend the same users as the original marriage purpose for the user, which leads to a contradiction between static initial data and dynamic behavior demand, resulting in inaccurate recommendations and reducing the customer matching effect of the marriage platform. SUMMARY

[0004] In order to solve the technical problem of poor customer matching effect of the marriage platform, the purpose of the present application is to provide a customer behavior analysis system for a marriage platform CRM, and the technical solution adopted is as follows:

[0005] The customer behavior analysis system for the marriage platform CRM comprises:

[0006] A customer information acquisition module is used to acquire portrait information of all customers in the marriage platform CRM system and interaction information of any target customer with other customers in each recommendation period; the portrait information at least includes marriage purpose;

[0007] An intelligent recommendation customer module is used to cluster customers based on a fuzzy clustering algorithm and the portrait information in the first recommendation period, to acquire the membership degree of the target customer in each cluster, and to acquire all recommended customers of the target customer according to the membership degree; in each of the remaining recommendation periods, the recommendation index of each recommended customer in the last recommendation period is acquired according to the interaction information of the target customer with each recommended customer in the adjacent last recommendation period, and the membership degree of the target customer in each cluster and the recommended customer are updated according to the recommendation index of all recommended customers in each cluster;

[0008] An image dynamic updating module is configured to obtain a purpose explicit index of the target customer according to a time sequence change of the membership of the target customer in each cluster at a current time, and adjust the marriage purpose of the target customer in combination with the marriage purpose of all customers in each cluster.

[0009] Further, the interaction information at least includes a browsing duration, an interaction frequency, and a single interaction duration.

[0010] Further, the method for obtaining the recommended customers includes:

[0011] The membership of the target customer in each cluster is multiplied by a preset multiple and then rounded to obtain a recommended number of customers in each cluster, and the recommended number of customers in each cluster is filtered as recommended customers of the target customer.

[0012] Further, the method for filtering the recommended number of customers in each cluster as recommended customers of the target customer includes:

[0013] In each cluster, the customers are sorted in descending order of the membership, and the recommended number of customers before sorting is taken as recommended customers.

[0014] Further, the method for obtaining the recommended index includes:

[0015] In a non-first recommended period, any recommended customer of the target customer in an adjacent previous recommended period is taken as an analyzed customer.

[0016] When the interaction frequency between the target customer and the analyzed customer is 0, a first parameter is obtained according to the browsing duration between the target customer and the analyzed customer, and the first parameter is taken as a recommended index of the analyzed customer.

[0017] When the interaction frequency between the target customer and the analyzed customer is greater than 0, a second parameter is obtained according to the interaction frequency and the single interaction duration between the target customer and the analyzed customer, and the second parameter is taken as a recommended index of the analyzed customer; wherein the second parameter is greater than the first parameter.

[0018] Further, the method for obtaining the second parameter includes:

[0019] A total interaction duration is obtained according to the interaction frequency and the single interaction duration, and a product of the total interaction duration and the interaction frequency is taken as a first interaction index.

[0020] In each non-first recommendation period, an interaction frequency change curve is fitted according to the total number of daily interactions between the target customer and the customer to be analyzed, and an interaction duration change curve is fitted according to the single interaction duration of the target customer and the customer to be analyzed each time of interaction; a second interaction index is obtained according to the slope of the interaction frequency change curve and the interaction duration change curve;

[0021] The first interaction index and the second interaction index are fused to obtain a second parameter.

[0022] Further, the method for updating the membership of the target customer in each cluster and the recommended customers comprises:

[0023] In each cluster, the sum of the recommendation indexes of all recommended customers is normalized to obtain a membership adjustment value, and the normalized value of the sum of the membership of the target customer in the cluster and the membership adjustment value is taken as the updated membership;

[0024] The updated membership of the target customer in each cluster is multiplied by a preset multiple and then rounded to obtain the updated recommendation number of customers in each cluster, and the updated recommendation number of customers in each cluster is filtered out as the updated recommended customers of the target customer.

[0025] Further, the method for obtaining the purpose explicit index comprises:

[0026] Until the current time, the membership of the target customer in each cluster in each recommendation period is taken as a data point to fit a time sequence change curve; the slope of each time sequence change curve is calculated, and the maximum slope is taken as a comparative change parameter;

[0027] In the most recent recommendation period at the current time, the cluster with the maximum membership of the target customer in all clusters is taken as a target cluster, and a highlight parameter of the target cluster is obtained according to the maximum membership and the deviation of the maximum membership from the rest of the memberships;

[0028] The slope of the time sequence change curve corresponding to the target cluster is taken as a change parameter, a purpose bias parameter is obtained according to the difference between the change parameter and the comparative change parameter, and the normalized value of the fusion result is taken as a purpose explicit index.

[0029] Further, the method for obtaining the highlight parameter comprises:

[0030] The difference between the maximum membership of the target customer in all clusters and the second largest membership is multiplied by the maximum membership, and the product is taken as the highlight parameter of the target cluster corresponding to the maximum membership.

[0031] Further, the method for adjusting the marriage purpose of the target customer comprises:

[0032] When the purpose clear index is greater than the preset threshold, the marriage purpose with the highest occurrence frequency in the marriage purposes of all the customers in the target cluster is taken as the changed marriage purpose of the target customer; when the purpose clear index is less than or equal to the preset threshold, the marriage purpose of the target customer is not updated.

[0033] The present application has the following beneficial effects:

[0034] The present application firstly acquires the portrait information of all the customers in the marriage platform CRM system and the interaction information of any target customer with other customers in each recommendation period, and prepares for the analysis of the subsequent customer behavior; then in the first recommendation period, the customers are clustered based on the fuzzy clustering algorithm and the portrait information, and all the recommended customers of the target customer are preliminarily acquired; further in each of the remaining recommendation periods, the potential expectation of the target customer is analyzed according to the interaction information of the target customer with each recommended customer in the adjacent last recommendation period, the recommendation index of each recommended customer in the last recommendation period is acquired, and the membership degree of the target customer in each cluster and the recommended customer are updated according to the recommendation index of all the recommended customers in each cluster, so as to continuously adjust the recommended candidates and gradually clarify the marriage purpose of the target customer; finally at the current time, the tendency of the target customer is clarified according to the time sequence change of the membership degree of the target customer in each cluster, the purpose clear index of the target customer is acquired, and the marriage purpose of the target customer is adjusted in combination with the marriage purpose of all the customers in each cluster. The present application analyzes the potential expectation of the target customer by analyzing the interaction behavior change of the target customer with the recommended customers of the platform in different recommendation periods, so as to continuously adjust the recommended candidates and gradually clarify the marriage purpose of the target customer, so as to improve the matching effect of the marriage platform. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0036] Figure 1 The system block diagram of the customer behavior analysis system for the marriage platform CRM provided by one embodiment of the present application;

[0037] Figure 2 The flowchart of the acquisition method of the purpose clear index provided by one embodiment of the present application. DETAILED DESCRIPTION

[0038] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the specific implementation, structure, features and effects of a customer behavior analysis system for a marriage platform CRM according to the present application are described in detail below in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0040] The specific scheme of a customer behavior analysis system for a marriage platform CRM provided by the present application is described in detail below in combination with the drawings.

[0041] Please refer to Figure 1 , which shows a system block diagram of a customer behavior analysis system for a marriage platform CRM provided by one embodiment of the present application, including a customer information acquisition module 101, an intelligent recommended customer module 102 and a portrait dynamic updating module 103.

[0042] Customer information acquisition module 101: used to acquire portrait information of all customers in the marriage platform CRM system, and interaction information of any target customer with other customers in each recommendation period; the portrait information at least includes marriage purpose.

[0043] In one embodiment of the present application, first, portrait information of each customer in the platform is collected based on the CRM system of the marriage platform, wherein the portrait information refers to personal information actively filled out or selected by the customer when registering on the marriage platform, including: basic information such as nickname, gender, age, zodiac and social account, etc.; education background such as education, college and major, etc.; professional information such as occupation and income, etc.; family information such as hometown, family members and asset information, etc.; living habits such as eating habits, work and rest time, exercise habits and smoking and drinking habits, etc.; interests and hobbies such as music, film, books and travel, etc.; marriage purpose such as quickly finding a love object, quickly finding a marriage object, ordinary friendship or long-term development, etc. one of the options; the implementer can also define the types and quantities of portrait information such as adding ideal type portrait information, but at least including marriage purpose to match ideal customers with the same target.

[0044] Considering that the marriage purpose of each client may change after contacting different recommended clients, such as gradually changing from ordinary friendship to intended love or marriage, the change of the marriage purpose may cause the subsequent recommended clients not to meet the intention of the target client, therefore, the embodiment of the present application sets a recommended period to periodically recommend matching clients to the target client, specifically, recommending clients to the target client at the beginning of each recommended period, and further analyzing the potential expectations of the target client according to the interaction behavior changes of the target client and the recommended clients of the marriage platform, so as to continuously adjust the recommended candidates, and gradually clarify the marriage purpose of the target client, thereby improving the matching effect of the marriage platform.

[0045] Therefore, in an embodiment of the present application, each week is specifically taken as a recommended period, and the CRM system of the marriage platform tracks the interaction information of each client with other clients in each recommended period, the interaction at least includes browsing, liking, commenting and sending information, and the obtained interaction information includes the browsing time length of the home page of other clients, the interaction times with other clients and the single interaction time length; the implementer can also customize the types and quantities of interaction information.

[0046] It should be noted that collecting the portrait information of the client and tracking the behavior of the client based on the CRM system is prior art, and will not be described in detail; the method of analyzing the behavior of each client in the marriage platform and changing the portrait is the same, and the embodiment of the present application will analyze and describe the behavior analysis and portrait change of a target client selected from all clients.

[0047] The intelligent recommended client module 102 is used for clustering the clients based on the fuzzy clustering algorithm and the portrait information in the first recommended period, obtaining the membership degree of the target client in each cluster, and obtaining all recommended clients of the target client according to the membership degree; in each of the remaining recommended periods, the interaction information of the target client with each recommended client in the adjacent last recommended period is obtained, the recommended index of each recommended client in the last recommended period is obtained, and the membership degree of the target client in each cluster and the recommended client are updated according to the recommended index of all recommended clients in each cluster.

[0048] Considering that the marriage purpose filled by the target client when initially registering is the real marriage intention at that time, the intended clients with similar or matching portraits can be directly recommended to the target client; considering that the portrait information and behavior mode of the clients in the marriage platform are diverse, the interaction tendency of the clients with similar portrait information is usually relatively strong, and the pairing success rate is also relatively high, and the fuzzy clustering algorithm can better adapt to the diversity of the clients, allowing a client to belong to multiple groups, thereby helping to capture the diversified intention demand of the clients and recommending the intended users to the clients.

[0049] Therefore, in one embodiment of the present application, firstly, the customers are clustered based on the fuzzy clustering algorithm and the customer portrait information in the first recommendation period, the membership degrees of the target customer in each cluster are obtained, and then all the recommended customers of the target customer are obtained according to the membership degrees; each cluster contains similar customer groups, and the membership degree of the customer in each cluster reflects the degree of belonging to the cluster, and recommending the customers in the customer group with high membership degree to the target customer will improve the marriage matching effect;

[0050] Specifically, firstly, each item of personal information in the customer portrait information is taken as a vector element to construct a multi-dimensional feature vector, then the multi-dimensional feature vector is taken as the customer label, the fuzzy C-means clustering algorithm is used to cluster all the customers, so that the customers with similar labels belong to the same cluster, wherein the number of clusters is set to the number of types of marriage purposes; then the cluster containing the target customer is selected for subsequent analysis, that is, the membership degree of the target customer relative to the cluster not containing the target customer is 0, and the subsequent recommended customers are all the customers in the cluster containing the target customer.

[0051] It should be noted that in one embodiment of the present application, the customers participating in the fuzzy clustering are the opposite-sex customers of the target customer, and the implementer can also adjust according to actual needs.

[0052] It should be noted that the construction of the multi-dimensional feature vector and the application of the fuzzy C-means clustering algorithm are both prior art and will not be described in detail; in other embodiments, the implementer can also use other fuzzy clustering algorithms, or can define the number of clusters by himself.

[0053] Preferably, in one embodiment of the present application, it is considered that the higher the membership degree of the target customer relative to the cluster, the better the effect of recommending the customers in the cluster to the target customer; and the value range of the membership degree is usually [0, 1], and the sum of the membership degrees of a customer relative to all clusters is 1, so the membership degree can be enlarged by a corresponding multiple and then rounded according to the number of recommended customers set in each recommendation period, and then the number of recommended customers of each cluster can be obtained, and then the corresponding customers are selected for recommendation; based on this, the method for obtaining the recommended customers comprises:

[0054] The membership degree of the target customer in each cluster is multiplied by a preset multiple and rounded to obtain the number of recommended customers in each cluster, and the recommended number of customers in each cluster is selected as the recommended customers of the target customer; wherein, in one preferred embodiment of the present application, in each cluster, the customers are sorted in descending order of membership degree, and the recommended number of customers before sorting is selected as the recommended customers.

[0055] As an example, the preset multiple is specifically set to 10, that is, the ideal number of recommended customers set in each recommendation period, and the implementer can also adjust it according to the actual application situation; the membership of the target customer in each cluster is multiplied by 10 and rounded up to obtain the number of recommended customers in each cluster; then the recommended customers with the top recommended number in each cluster can be screened out according to the membership ranking, and the union of the recommended customers in all clusters is the recommended customer set of the target customer.

[0056] In other examples, the implementer can also measure the similarity between the multi-dimensional feature vector of each customer in each cluster and the multi-dimensional feature vector corresponding to the target customer, for example, using Euclidean distance, DTW distance, etc. to measure the difference, and then sorting the customers in descending order according to the difference, and then selecting the recommended number of recommended customers similar to the target customer.

[0057] So far, the customer recommendation for the target customer in the first recommendation period has been completed; further, the recommendation satisfaction of the target customer to each recommended customer, that is, the recommendation index, can be analyzed and evaluated by combining the interaction information of the target customer and all recommended customers, so as to analyze and evaluate the intention of the customer through the behavior information, and then continuously adjust the recommended customers to clarify the marriage intention of the target customer.

[0058] Therefore, in the second recommendation period, the embodiment of the present application first obtains the recommendation index of each recommended customer in the previous recommendation period according to the interaction information of the target customer and each recommended customer in the adjacent previous recommendation period, and further updates the membership of the target customer in each cluster and the recommended customers according to the recommendation index of all recommended customers in each cluster; then by repeatedly updating the above process in each new recommendation period, the membership of the target customer in each cluster and the recommended customers are iteratively updated to gradually clarify the marriage intention of the target customer.

[0059] It should be noted that the iterative updating process in each non-first recommendation period is consistent, and only any non-first recommendation period is taken as an example for analysis and description, the recommendation index of each recommended customer in the previous recommendation period is obtained, and the membership of the target customer in each cluster and the recommended customers are updated according to the recommendation index of all recommended customers in each cluster.

[0060] Preferably, in one embodiment of the present application, any recommended customer of the target customer in the last recommended period is first taken as an analyzed customer, and the analyzed customer is taken as an example for analysis; considering that if the target customer only browses and does not interact with the analyzed customer, it means that the target customer is less interested in the analyzed customer, and the browsing time also reflects the interest degree, the shorter the browsing time, the lower the matching intention, and the lower the recommended index of the analyzed customer recommended in the last recommended period; when the target customer interacts with the analyzed customer, the more the interaction times and the longer the interaction time, the more interested the target customer is in the analyzed customer, and the greater the recommended index; but the recommended index when interacting should always be greater than the recommended index when not interacting; based on this, the method for obtaining the recommended index comprises:

[0061] When the interaction times of the target customer and the analyzed customer is 0, a first parameter is obtained according to the browsing time of the target customer to the analyzed customer, and the first parameter is taken as the recommended index of the analyzed customer;

[0062] When the interaction times of the target customer and the analyzed customer is greater than 0, a second parameter is obtained according to the interaction times and the single interaction time of the target customer and the analyzed customer, and the second parameter is taken as the recommended index of the analyzed customer; wherein the second parameter is greater than the first parameter.

[0063] In one preferred embodiment of the present application, considering that the interaction times and the total interaction time in the recommended period can help to evaluate the interest of the target customer and the analyzed customer, a first interaction index can be obtained to help to evaluate the recommended index or the second parameter; and considering that in each recommended period, as the interaction deepens, the target customer and the analyzed customer understand each other further, at this time the target customer may think that the analyzed customer does not meet his expectations, and then reduces the interaction times and the interaction time, a second interaction index can be obtained according to the change of the interaction; and then the two interaction indexes are integrated to evaluate the recommended index or the second parameter of the analyzed customer; therefore, the method for obtaining the second parameter comprises:

[0064] The total interaction time is obtained according to the interaction times and the single interaction time, and the product of the total interaction time and the interaction times is taken as the first interaction index;

[0065] In each non-first recommended period, an interaction frequency change curve is fitted according to the total interaction times of the target customer and the analyzed customer per day, and an interaction time change curve is fitted according to the single interaction time of the target customer and the analyzed customer each time; the second interaction index is obtained according to the slope of the interaction frequency change curve and the interaction time change curve;

[0066] The first interaction index and the second interaction index are fused to obtain the second parameter.

[0067] As an example, first set a limit between interaction and non-interaction, such as 0.3, so that the value range of the first parameter is [0, 0.3], and the value range of the second parameter is (0.3, 1];

[0068] When the number of interactions between the target customer and the customer to be analyzed is 0, the browsing time of the target customer to the customer to be analyzed is linearly normalized, and the normalized value multiplied by 0.3 is taken as the first parameter. The implementer can also directly map to obtain the first parameter, and then obtain the recommendation index of the customer to be analyzed;

[0069] When the number of interactions between the target customer and the customer to be analyzed is greater than 0, the single interaction time of each interaction is accumulated to obtain the total interaction time, and then the first interaction index is obtained. Further, the interaction frequency curve and the interaction time curve are fitted, the slope of the two curves is obtained based on the two-point formula, and the sum of the two slopes is taken as the second interaction index. Finally, the first interaction index and the second interaction index are added and fused, and the product is mapped into (0.3, 1] to obtain the second parameter, and then the recommendation index of the customer to be analyzed is obtained.

[0070] It should be noted that mapping, fitting of change curve and calculation of slope are all prior art; in other examples, the implementer can also adjust the limit, and can also fuse the first interaction index and the second interaction index through multiplication or weighted summation, etc. Basic mathematical operations, or other normalization or mapping means can also be used, which will not be described here.

[0071] Change the customer to be analyzed, and the recommendation index of each recommended customer in the last recommendation period can be obtained; further, considering that the higher the recommendation index of the recommended customer in each cluster, the higher the intention degree of the target customer to the customer in the cluster, the membership of the target customer to the cluster should also be higher, then the membership of the target customer to each cluster and the recommended customer can be updated according to the recommendation index of all recommended customers in each cluster.

[0072] Preferably, in an embodiment of the present application, considering that the recommendation index of the recommended customer of the target customer in each cluster can be used as a further reference for the membership, first, the membership adjustment value can be evaluated based on the recommendation index of all recommended customers, and then the membership of the target customer to the cluster is updated. Adjust; then update the recommended customer; based on this, the method for updating the membership of the target customer to each cluster and the recommended customer comprises:

[0073] In each cluster, the sum of the recommendation indexes of all recommended customers is normalized to obtain the membership adjustment value, and the normalized value of the sum of the membership of the target customer in the cluster and the membership adjustment value is taken as the updated membership.

[0074] The updated membership degree of the target customer in each cluster is multiplied by a preset multiple and then rounded to obtain the updated recommendation number of the customer in each cluster. The updated recommendation number of customers in each cluster is then selected as the updated recommended customers of the target customer.

[0075] As an example, in each non-first recommendation period, the sum of recommendations from all recommended customers in each cluster is first linearly normalized to obtain the membership adjustment value. It should be noted that all recommended customers participating in the summation of recommendation indices in each cluster should not only include the recommended customers selected in each cluster, but should include the recommended customers in the union of all recommended customers in the above clusters.

[0076] Then, the membership degree of the target customer obtained in the previous recommendation period is added to the membership degree adjustment value, and the sum is used as the initial updated membership degree of the target customer in each cluster. Then, the initial updated membership degree is normalized to obtain the updated membership degree. Furthermore, based on the same acquisition method as the recommended customers in the first recommendation period, the updated recommended customers of the target customer in each cluster can be obtained. The specific steps will not be repeated.

[0077] The specific method for normalizing the initial update membership is as follows: take the initial update membership of the target customer in each cluster as the numerator, take the sum of the initial update membership of the target customer in all clusters as the denominator, and take the ratio of the fractions as the normalized value.

[0078] This allows for continuous updates of the target customer's recommended customers within each recommendation cycle.

[0079] Profile Dynamic Update Module 103: Used to obtain the target customer's purpose clarification index based on the temporal changes in the target customer's membership degree in each cluster at the current moment, and adjust the target customer's marriage and love purpose in combination with the marriage and love purpose of all customers in each cluster.

[0080] Considering the changes in the membership degree of target customers in each cluster at the current moment, it can reflect the changes in their intentions towards that cluster, determine the target customer's purpose clarity index, thereby helping to clarify the target customer's intended group or purpose, and then update and adjust the target customer's purpose for marriage and love based on the marriage and love purposes of customers within the intended group.

[0081] Preferably, in one embodiment of the present invention, the method for obtaining the purpose-defined index includes:

[0082] Please see Figure 2 The diagram illustrates a flowchart of a method for obtaining a purpose-defined index according to an embodiment of the present invention, specifically including:

[0083] Step S201, until the current time, the membership of the target customer in each cluster in each recommendation period is taken as a data point, and a time sequence change curve is fitted; the slope of each time sequence change curve is calculated, and the maximum slope is taken as a comparison change parameter.

[0084] It is considered that the higher the membership of the target customer relative to the cluster becomes with the update of the recommendation period, the more explicit the intention of the target user to the cluster is, and the more inclined to the group, and fitting the time sequence change curve of the membership and then calculating the slope can help to evaluate the change of the membership.

[0085] Therefore, as an example, taking any cluster as an example, the membership of the target customer in the cluster in each recommendation period is taken as a data point, and then mapped to the corresponding timestamp, and the time sequence change curve of the membership is fitted; then the slope of each cluster corresponding to the time sequence change curve is calculated based on two-point formula, and the maximum slope is selected as the comparison change parameter; the comparison change parameter reflects the inclination of the target customer to the cluster, and also reflects the explicitness of the purpose.

[0086] It should be noted that the fitting of the time sequence change curve and the acquisition of the slope are prior art and will not be described again.

[0087] Step S202, in the latest recommendation period at the current time, the cluster with the maximum membership of the target customer in all clusters is taken as a target cluster, and the highlight parameter of the target cluster is acquired according to the maximum membership and the deviation of the maximum membership relative to the rest of the membership.

[0088] It is considered that in the latest recommendation period at the current time, the cluster with the maximum membership after the update of the membership of the target customer relative to each cluster will reflect the group inclination of the target customer at the latest time of the recommendation update, which can also help to evaluate the purpose of the target customer; and it is considered that in the latest recommendation period at the current time, if the membership of the target customer in a cluster is higher relative to the membership in other clusters, the intention of the target customer to the cluster is more explicit, which can help to evaluate the highlight coefficient of the cluster, and then facilitate the subsequent comprehensive evaluation of the explicitness index of the purpose of the target customer.

[0089] In a preferred embodiment of the present application, it is considered that the cluster corresponding to the maximum membership is the target cluster, and the greater the difference between the maximum membership and the second largest membership of the target customer in all clusters is, the more explicit the matching intention of the target customer to the customer group in the target cluster is, and then the highlight coefficient of the target cluster can be evaluated; the method for acquiring the highlight parameter comprises:

[0090] The difference between the maximum membership and the second largest membership of the target customer in all clusters is multiplied by the maximum membership, and the product is taken as the highlight parameter of the target cluster corresponding to the maximum membership.

[0091] In step S203, the slope of the time sequence change curve corresponding to the target cluster is taken as a change parameter, a purpose bias parameter is obtained according to the difference between the change parameter and a contrast change parameter, and a fusion result is obtained by fusing the highlight parameter and the purpose bias parameter, and a normalized value of the fusion result is taken as a purpose explicit index.

[0092] It is considered that the cluster corresponding to the time sequence change curve corresponding to the maximum slope may not be the target cluster, but when the slope of the time sequence change curve corresponding to the target cluster is closer to the maximum slope, it is indicated that the pairing intention of the target customer to the customers in the target cluster is also more obvious, so that the purpose bias parameter of the target customer to the target cluster can be obtained, and then the purpose explicit index of the target customer to the pairing of the customers in the target cluster can be evaluated by fusing the purpose bias parameter and the highlight parameter.

[0093] As an example, the change parameter is taken as a numerator, the contrast change parameter is taken as a denominator, and a fractional ratio is taken as a purpose bias parameter, and when the ratio is closer to 1, the purpose bias parameter is larger. Then, the purpose bias parameter and the highlight parameter are multiplied to be fused, and the product is linearly normalized to obtain the purpose explicit index.

[0094] In other examples, the implementer can also obtain the purpose bias parameter by negatively correlating the difference between the contrast change parameter and the change parameter, such as taking an inverse operation, and the larger the inverse, the larger the obtained purpose bias parameter. Other normalization methods can also be used, which are not described herein.

[0095] After obtaining the purpose explicit index of the target customer, the marriage intention of the target customer can be further evaluated, and the customer can be reminded to change the marriage target to improve the pairing effect of the marriage platform.

[0096] Preferably, in an embodiment of the present application, the method for adjusting the marriage purpose of the target customer comprises:

[0097] When the purpose explicit index is greater than a preset threshold, the marriage purpose with the highest appearance frequency in the marriage purposes of all the customers in the target cluster is taken as the changed marriage purpose of the target customer; and when the purpose explicit index is less than or equal to the preset threshold, the marriage purpose of the target customer is not updated.

[0098] Wherein, in one example of the present application, the preset threshold is set to 0.7, and the implementer can also customize it; when the purpose clear index is greater than 0.7, it is determined that the marriage purpose of the target customer is gradually clear, and the marriage platform sends a reminder to the target customer, such as: "the platform analyzes your behavior and identifies that your current marriage purpose may deviate from your initial marriage purpose, and hopes that you can modify it in your personal information to facilitate the platform to recommend more suitable matching candidates for you"; after the target customer modifies the marriage purpose, the above operation is repeated to re-recommend customers for it until the target customer finds the ideal customer; when the purpose clear index is less than or equal to 0.7, continue to update the recommended customers for the target customer in the new recommendation period until the target customer finds the ideal customer.

[0099] To sum up, the present application first acquires the portrait information of all customers and the interaction information of the target customer with other customers in each recommendation period; then in the first recommendation period, the customers are clustered based on the fuzzy clustering algorithm and the portrait information to obtain all the recommended customers of the target customer; further in each of the remaining recommendation periods, the membership degree of the target customer in each cluster and the recommended customers are updated according to the interaction information of the target customer with each recommended customer; finally, at the current time, the purpose clear index of the target customer is obtained according to the time sequence change of the membership degree of the target customer in each cluster, and the marriage purpose of the target customer is adjusted. The present application analyzes the interaction behavior change of the target customer with the recommended customers of the platform in different recommendation periods, analyzes the potential expectations of the target customer, and thus continuously adjusts the recommended candidates, and gradually clarifies the marriage purpose of the target customer, so as to improve the matching effect of the marriage platform.

[0100] It should be noted that the above-mentioned embodiments of the present application are only for description, not representing the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or may be advantageous.

[0101] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other, and each embodiment mainly describes the difference from other embodiments.

Claims

1. A customer behavior analysis system for a marriage platform CRM, characterized in that, The system comprises: a customer information collection module, configured to acquire portrait information of all customers in a marriage platform CRM system and interaction information of any target customer with other customers in each recommendation period, wherein the portrait information at least comprises a marriage purpose; an intelligent recommendation customer module, configured to, in a first recommendation period, cluster the customers based on a fuzzy clustering algorithm and the portrait information, acquire a membership degree of the target customer in each cluster, and acquire all recommended customers of the target customer according to the membership degree; and in each remaining recommendation period, acquire a recommendation index of each recommended customer in a previous recommendation period according to the interaction information of the target customer with each recommended customer in the previous recommendation period, and update the membership degree of the target customer in each cluster and the recommended customers according to the recommendation index of all recommended customers in each cluster; a portrait dynamic updating module, configured to, at a current time, acquire a purpose clear index of the target customer according to a time sequence change of the membership degree of the target customer in each cluster, and adjust the marriage purpose of the target customer in combination with the marriage purpose of all customers in each cluster; the acquisition method of the purpose clear index comprises: until the current time, taking the membership degree of the target customer in each cluster in each recommendation period as a data point, fitting a time sequence change curve; calculating a slope of each time sequence change curve, and taking a maximum slope as a comparative change parameter; in a latest recommendation period at the current time, taking a cluster with the maximum membership degree of the target customer in all clusters as a target cluster, and acquiring a highlight parameter of the target cluster according to the maximum membership degree and a deviation of the maximum membership degree relative to the remaining membership degrees; taking the slope of the time sequence change curve corresponding to the target cluster as a change parameter, acquiring a purpose bias parameter according to a difference between the change parameter and the comparative change parameter; and fusing the highlight parameter and the purpose bias parameter, taking a normalized value of a fusion result as the purpose clear index; the acquisition method of the highlight parameter comprises: multiplying a difference between the maximum membership degree of the target customer in all clusters and a second largest membership degree by the maximum membership degree, and taking a product as the highlight parameter of the target cluster corresponding to the maximum membership degree. 2.The customer behavior analysis system for a marriage platform CRM according to claim 1, wherein, The interaction information at least comprises a browsing time length, an interaction frequency and a single interaction time length. 3.The customer behavior analysis system for a marriage-oriented platform CRM according to claim 1, wherein, the acquisition method of the recommended customers comprises: multiplying the membership degree of the target customer in each cluster by a preset multiple, obtaining a recommended number of customers in each cluster, and selecting the recommended number of customers in each cluster as the recommended customers of the target customer.

4. The customer behavior analysis system for a marriage-oriented platform CRM according to claim 3, characterized in that, the method of selecting the recommended number of customers in each cluster as the recommended customers of the target customer comprises: in each cluster, sorting the customers in descending order of the membership degree, and selecting the recommended number of customers in the sorted order as the recommended customers. 5.The customer behavior analysis system for a marriage-oriented platform CRM according to claim 2, wherein, the acquisition method of the recommendation index comprises: in a non-first recommendation period, taking any recommended customer of the target customer in a previous recommendation period as an analyzed customer; when the interaction frequency of the target customer with the analyzed customer is 0, acquiring a first parameter according to the browsing time length of the target customer with the analyzed customer, and taking the first parameter as the recommendation index of the analyzed customer. When the number of interactions between the target client and the client to be analyzed is greater than 0, a second parameter is obtained according to the number of interactions between the target client and the client to be analyzed and the single interaction duration, and the second parameter is used as a recommended index of the client to be analyzed; wherein the second parameter is greater than the first parameter.

6. The customer behavior analysis system for a marriage-oriented platform CRM according to claim 5, characterized in that, The method for obtaining the second parameter comprises: a total interaction duration is obtained according to the number of interactions and the single interaction duration, and a product of the total interaction duration and the number of interactions is used as a first interaction index; in each non-first recommended period, an interaction frequency change curve is fitted according to the total number of interactions between the target client and the client to be analyzed per day, and an interaction duration change curve is fitted according to the single interaction duration of the target client and the client to be analyzed each time; a second interaction index is obtained according to the slope of the interaction frequency change curve and the interaction duration change curve; the first interaction index and the second interaction index are fused to obtain the second parameter.

7. The customer behavior analysis system for a marriage platform CRM according to claim 1, characterized in that, The method for updating the membership of the target client in each cluster and the recommended clients comprises: in each cluster, the sum of the recommended indexes of all recommended clients is normalized to obtain a membership adjustment value, and the normalized value of the sum of the membership of the target client in the cluster and the membership adjustment value is used as the updated membership; the updated membership of the target client in each cluster is multiplied by a preset multiple and then rounded to obtain the updated recommended number of clients in each cluster, and the updated recommended number of clients in each cluster is filtered out as the updated recommended clients of the target client. 8.The customer behavior analysis system for a marriage platform CRM according to claim 1, wherein, The method for adjusting the marriage purpose of the target client comprises: when the purpose clear index is greater than a preset threshold, the marriage purpose with the highest appearance frequency in the marriage purpose of all clients in the target cluster is used as the changed marriage purpose of the target client; when the purpose clear index is less than or equal to the preset threshold, the marriage purpose of the target client is not updated.

Images