Investigation and promotion method and system based on shared information

Abstract

The invention relates to the technical field of product investigation and promotion, and discloses an investigation and promotion method and system based on shared information, and the method comprises the steps: obtaining session shared information of a user; identifying a behavior pattern of a user based on the session sharing information; the method specifically comprises the steps of identifying a browsing behavior mode and a purchasing behavior mode of a user; obtaining a high-intention product of the user based on the browsing behavior mode; obtaining an associated intention product of the user based on the purchase behavior mode; and carrying out product popularization on the user based on the high-intention product and the associated intention product. According to the method, deep user behavior analysis can be provided, so that the product recommendation accuracy is improved, and product popularization strategy optimization is facilitated.

Description

Technical Field

[0001] This application relates to the technical field of product investigation and promotion, and particularly to a method and system for investigation and promotion based on shared information. Background Art

[0002] By analyzing users' purchase records and browsing behaviors on the platform, e-commerce platforms can provide personalized recommendation services for users. However, there are still many deficiencies in the current industry in terms of product investigation and promotion. Although existing data analysis methods can uncover some superficial information, they are far from sufficient for a deep understanding of user behavior patterns. When analyzing users' browsing and purchase behaviors, only some simple associations can be found, and it is difficult to uncover deep-seated and non-intuitive product correlations. In terms of product promotion, the current personalized recommendation algorithms still need to be optimized. Most existing recommendation systems are based on simple collaborative filtering algorithms or content-based recommendation algorithms, and these algorithms have limitations in dealing with complex user behaviors and diverse product requirements. They can often only recommend similar products based on users' historical behaviors and cannot fully consider users' real-time needs and scenario changes.

[0003] For example, the Chinese patent with the authorization announcement number CN116823382B discloses a product promotion method based on big data, including: determining the product identifier of the product to be promoted; determining a target product set according to the product identifier, and obtaining a number of target customers corresponding to the target product set; the target product set includes a number of target products related to the product identifier of the product to be promoted; obtaining the consumption behavior information of each target customer, and determining the consumption label of the target customer according to the consumption behavior information; generating promotion information according to the consumption label and the product parameters of the product to be promoted; and sending the promotion information to the corresponding target customer. Through the product promotion method given by this invention, it is possible to determine the consumption label of the target customer based on the consumption behavior of the target customer and generate promotion information in combination with the product parameters of the product to be promoted. At the same time, since the promotion information has a high fit with the consumption behavior of the target customer, the product promotion effect is improved.

[0004] As disclosed in the Chinese patent application with the publication number CN106408366A, a product promotion method and a product promotion system are disclosed. The product promotion method includes the following steps: H. The user terminal receives a first retrieval instruction from a consumer, and the user terminal transmits the first retrieval instruction to the service background. The service background transmits the preliminary data information of the product to the user terminal. After receiving the preliminary data information, the user terminal displays it to the consumer; K. The user terminal receives a second retrieval instruction from the consumer, and the user terminal transmits the second retrieval instruction to the service background. The service background transmits the production data information of the product to the user terminal. After receiving the production data information, the user terminal displays it to the consumer. The invention can more comprehensively display the information of the product during the production process to consumers, improve the comprehensiveness and accuracy of product information, and provide valuable reference for consumers.

[0005] The above prior arts all have the problems raised in this background art: when analyzing users' browsing and purchasing behaviors, it is difficult to dig out deep-level and non-intuitive product correlations.

[0006] The information disclosed in this background art section is only intended to increase the understanding of the overall background of the present application, and should not be regarded as an admission or any form of implication that this information constitutes the prior art already known to those of ordinary skill in the art. Summary of the Invention

[0007] The technical problem to be solved by the present application is to overcome the defects of the prior art and provide a survey and promotion method and system based on shared information, which provides in-depth user behavior analysis, thereby improving the accuracy of product recommendation.

[0008] To solve the above technical problem, the present application provides the following technical solutions:

[0009] On the one hand, the present application provides a survey and promotion method based on shared information, including the following steps:

[0010] Obtain the session shared information of the user;

[0011] Identify the behavior pattern of the user based on the session shared information; specifically, it includes identifying the browsing behavior pattern and purchasing behavior pattern of the user;

[0012] Obtain the high-intention products of the user based on the browsing behavior pattern; obtain the associated intention products of the user based on the purchasing behavior pattern;

[0013] Promote products to the user based on the high-intention products and associated intention products.

[0014] As a preferred solution of the survey and promotion method based on shared information of the present application, wherein: the method for identifying the browsing behavior pattern of the user is as follows:

[0015] Extract each product browsing record of the user for the same type of product from the session shared information;

[0016] Count the types of relevant pages of the same type of product, and assign a number to each type of relevant page;

[0017] Based on the product browsing record and the numbers of relevant pages, establish a browsing feature sequence and a browsing order matrix to record the user's browsing behavior pattern.

[0018] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: the browsing feature sequence is used to record the user's attention to each type of relevant page; the browsing feature sequence is a vector with a length of N; wherein, N is the number of types of relevant pages of the same type of product; each element in the browsing feature sequence corresponds to a type of relevant page;

[0019] The method for establishing the browsing feature sequence is as follows:

[0020] Sort the N types of relevant pages, and record the numbers of relevant pages based on the sorting;

[0021] Respectively count the total browsing duration of each type of relevant page in all product browsing records, and calculate the total browsing duration of all relevant pages;

[0022] Calculate the element value corresponding to each number; form the browsing feature sequence with all element values; the element value corresponding to any number is the ratio of the total browsing duration of the relevant page corresponding to the number to the total browsing duration of all relevant pages.

[0023] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: the browsing order matrix is used to record the browsing order of the user for different relevant pages; the method for establishing the browsing order matrix is as follows:

[0024] Sort the numbers of the N types of relevant pages horizontally and vertically to form a browsing order matrix with N rows and N columns, and initialize each element in the browsing order matrix;

[0025] Based on the product browsing record, count the user's browsing path and the transition probability of each browsing path; any browsing path includes two relevant pages, namely the first page and the second page;

[0026] Let represent a browsing path where the first page is the i-th type of relevant page and the second page is the j-th type of relevant page, then the transition probability of the browsing path is denoted as , indicating the probability that the user browses the j-th type of relevant page after browsing the i-th type of relevant page;

[0027] Assign values to the elements in the browsing order matrix based on the transition probability of each browsing path; among them, the element value in the i-th row and j-th column of the browsing order matrix is .

[0028] As a preferred solution of the investigation and promotion method based on shared information described in this application, among them: The recognition of the user's browsing behavior pattern further includes dynamically correcting the browsing order matrix, specifically including:

[0029] Identify the user's dependent browsing path; set a transition probability threshold, and mark the browsing path with a transition probability greater than the transition probability threshold as a dependent browsing path;

[0030] If it is detected that the user is browsing the first page of the dependent browsing path, then select a page from the relevant pages not included in the dependent browsing path as the third page; push the third page to the user through a pop-up window;

[0031] Collect the operation feedback of the user on the pop-up window push; the operation feedback includes explicit positive feedback, implicit positive feedback, and negative feedback;

[0032] Correct the browsing order matrix based on the operation feedback.

[0033] As a preferred solution of the investigation and promotion method based on shared information described in this application, among them: If the user jumps and browses the third page through the pop-up window, the operation feedback is explicit positive feedback; if the user closes the pop-up window and manually jumps and browses the third page when finishing browsing the first page, the operation feedback is implicit positive feedback; if the user closes the pop-up window and does not jump to the third page when finishing browsing the first page, the operation feedback is negative feedback;

[0034] Correcting the browsing order matrix based on the operation feedback specifically includes:

[0035] Let the first page of the dependent browsing path be the p-th type of relevant page, the second page of the dependent browsing path be the q-th type of relevant page, and the third page be the r-th type of relevant page; if the user's operation feedback is explicit positive feedback, then the transition probability of the browsing path of decrease , and increase the transition probability of the browsing path of increase ; is the preset first correction amount; if the user's operation feedback is implicit positive feedback, then the transition probability of the browsing path of decrease , and the transition probability of the browsing path of Increase ; is a preset second correction amount; if the user's operation feedback is negative feedback, then keep and unchanged.

[0036] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: The method for identifying the purchase behavior pattern of a user is as follows:

[0037] Extract the user's product purchase records from the session shared information, and extract each product purchased by the user from the product purchase records; assign a product number to any product purchased by the user;

[0038] Record the purchase time of each product purchase by the user and the corresponding product number, and establish a purchase time series of the user; each position in the purchase time series corresponds to a time point; if the user does not purchase a product at any time point, the position corresponding to the time point is empty; if the user purchases a product at any time point, the element value at the position corresponding to the time point is the product number;

[0039] Set the length and sliding step of the sliding window; intercept the purchase time series based on the sliding window and the sliding step; create a purchase event based on each intercepted sequence segment;

[0040] Based on the product number and the purchase event, establish a purchase event table to record the purchase behavior pattern of the user.

[0041] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: Any column in the purchase event table corresponds to the product number of a product; any row in the purchase event table corresponds to a purchase event;

[0042] The method for establishing a purchase event table is as follows:

[0043] Create an empty purchase event table based on the number of product numbers and the number of purchase events;

[0044] Assign values to the empty purchase event table based on the product numbers included in each purchase event, specifically as follows: In any purchase event, the cell at the corresponding position in the purchase event table is assigned a value of 1 for any product purchased by the user, and the cell at the corresponding position in the purchase event table is assigned a value of 0 for any product not purchased by the user.

[0045] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: The method for obtaining the high-intent products of the user based on the browsing behavior pattern is as follows:

[0046] Obtain the product browsing records and product purchase records of other users;

[0047] Based on the product browsing records and product purchase records, filter out other users who have browsed and purchased the same type of product, and mark them as reference users;

[0048] Establish the browsing feature sequence and browsing order matrix for each reference user;

[0049] For any reference user, calculate the similarity between the user and the reference user's browsing feature sequences and the similarity between the browsing order matrices and sum them to obtain the similarity between the user and the reference user;

[0050] Set a similarity threshold; mark the reference users whose similarity with the user is higher than the similarity threshold as similar users;

[0051] Based on the product purchase records, extract the specific models of the same type of products purchased by each similar user; count the number of purchases of each specific model of product, and mark the m products with the most purchases as the high-intent products.

[0052] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: the method for obtaining the associated intent products of the user based on the purchase behavior pattern is as follows:

[0053] Set an association threshold and a purchase frequency threshold;

[0054] Combine any two products in the purchase event table whose purchase frequencies are both greater than the purchase frequency threshold to obtain a first association group; any first association group contains two products with different product numbers;

[0055] Count the association degree of each first association group, and delete the first association groups with an association degree less than the association threshold; the association degree of the first association group is the number of times the first association group appears in the purchase event table divided by the total number of purchase events;

[0056] Combine any two first association groups and remove duplicates to obtain a second association group; count the association degree of each second association group, and delete the second association groups with an association degree less than the association threshold; the association degree of the second association group is the number of times the second association group appears in the purchase event table divided by the total number of purchase events; save the first association group and the second association group;

[0057] Based on the product purchase records, extract the list of all purchased products of the user in the most recent n days; n is a positive integer; combine the first association group and the second association group to obtain the associated intent products of the user.

[0058] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: combining the first association group and the second association group to obtain the user's associated intended products specifically includes:

[0059] S101: If there is a second association group that contains any two purchased products, and there is a product that the user has not purchased in the second association group, then mark the product that the user has not purchased in the second association group as an associated intended product;

[0060] S102: If no associated intended product is obtained in S101, then obtain the associated intended product through the first association group, specifically as follows:

[0061] If there is a first association group that contains any one purchased product, and there is a product that the user has not purchased in the first association group, then mark the product that the user has not purchased in the first association group as an associated intended product;

[0062] Sort all the associated intended products based on the association degree of the corresponding first association group, and retain the associated intended products corresponding to the M first association groups with the largest association degree.

[0063] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: promoting products to users based on the high-intention products and associated intended products includes regular promotion; the method of the regular promotion includes:

[0064] Real-time detect the browsing behavior and purchasing behavior of the user based on the session shared information; when it is detected that the user has completed browsing any type of product and it is not detected that the user has purchased the same type of product, then obtain the high-intention products of the user and recommend high-intention products of the same type as the browsed product to the user;

[0065] When it is detected that the user has purchased a certain product, obtain the associated intended products of the user and recommend the associated intended products that are in the same second association group or first association group as the purchased product to the user.

[0066] As a preferred solution of the investigation and promotion method based on shared information described in this application, wherein: promoting products to users based on the high-intention products and associated intended products also includes scenario linkage promotion; the method of the scenario linkage promotion is as follows:

[0067] Obtain the high-intention products and associated intended products of the user and form a product recommendation list;

[0068] Extract the device information of the user from the session shared information; determine the common geographical location of the user device based on the device information; combine the product purchase records and calculate the purchase probability of each product in the product recommendation list at each common geographical location;

[0069] Mark the common geographical locations with the highest purchase probabilities as the associated geographical locations of the corresponding products;

[0070] If it is detected that the user's device is at the associated geographical location of any product in the product recommendation list, recommend the corresponding product to the user.

[0071] In a second aspect, the present application provides a survey and promotion system based on shared information, including an information acquisition module, a behavior recognition module, a product survey module, and a product promotion module; where:

[0072] The information acquisition module is used to acquire the session shared information of the user;

[0073] The behavior recognition module identifies the user's behavior pattern based on the session shared information;

[0074] The product survey module obtains the user's high-intent products based on the browsing behavior pattern and obtains the user's associated intent products based on the purchase behavior pattern;

[0075] The product promotion module is used to promote products to the user, including regular promotion and scenario-linked promotion for the user.

[0076] Compared with the prior art, the beneficial effects achieved by the present application are as follows:

[0077] By acquiring the session shared information including user identification, product browsing records, product purchase records, and device and environment information, the user's behavior can be comprehensively and deeply analyzed, the user's attention degree and browsing order of the product page can be grasped, the non-intuitive correlation between products can be discovered, and strong support can be provided for deeply understanding the user's consumption habits and preferences. Analyzing the user's browsing and purchase behavior patterns can respectively obtain the user's high-intent products and associated intent products, making the product recommendation more in line with the user's needs and improving the accuracy of product promotion and the probability of the user purchasing related products. The product promotion method adopted by the present application covers regular promotion and scenario-linked promotion. It can not only recommend products in real time according to the user's browsing and purchase behaviors, but also actively recommend products to the user in a suitable scenario, enhancing the product promotion effect. Description of the Drawings

[0078] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts. Among them:

[0079] Figure 1Flowchart of the investigation and promotion method based on shared information provided by this application;

[0080] Figure 2 Structural schematic diagram of the investigation and promotion system based on shared information provided by this application. Detailed implementation manners

[0081] The technical solution of this application will be described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of this application and the specific features in the embodiments are detailed descriptions of the technical solution of this application, rather than limitations on the technical solution of this application. Without conflict, the technical features in the embodiments of this application and the embodiments can be combined with each other.

[0082] Embodiment 1

[0083] This embodiment introduces an investigation and promotion method based on shared information. Referring to Figure 1 , the method includes the following steps:

[0084] Obtain the session shared information of the user;

[0085] The session shared information of the user includes user identification, product browsing records, product purchase records, device and environment information;

[0086] The user identification is used to identify the user's identity. Common user identifications include user ID, login account, mobile phone number, email, etc. The user identification can associate the browsing and purchasing behaviors of the same user at different times and on different pages. The product browsing records include the browsing records of any type of product, specifically including parameters such as the page URL (page address) browsed by the user and the page stay duration. By analyzing the page stay duration, the degree of interest of the user in different page contents of any type of product can be judged. The product purchase records include parameters such as the product information, quantity, and purchase time purchased by the user. The product purchase records can directly reflect the user's consumption preferences and provide an important basis for mining associated intended products. The device and environment information includes the type of device used by the user (such as mobile phone, computer, tablet), device IP address, device geographical location, etc. These information helps to understand the user's device usage and purchase scenarios, and targeted product promotion can be carried out according to different devices and geographical locations.

[0087] Session shared information refers to the data shared by the server among multiple requests to maintain the user state. In many application scenarios, such as adding items to the shopping cart, accessing product pages, and user login and logout on e-commerce websites, the server needs to be able to identify that these requests come from the same session (i.e., user conversation). When a user first accesses the server, the server creates a new Session and generates a unique Session ID. Subsequently, the server passes this Session ID to the client, usually by setting a Cookie. The client will carry this Session ID in every subsequent request, so that the server can identify that this is a request from the same session. The server then looks up the corresponding Session data based on the received Session ID to restore the user state. Session data is generally stored on the server side, while the client only saves the Session ID, which helps to protect the security of sensitive information. Existing technologies include various ways to obtain session shared information, such as through Cookies, through databases, and through cache servers.

[0088] Identify the user's behavior pattern based on the session shared information; specifically, it includes identifying the user's browsing behavior pattern and purchasing behavior pattern;

[0089] The browsing behavior pattern represents the browsing order and degree of attention of the user to the relevant pages of a certain type of product when browsing that type of product. The user's browsing behavior is often sequential (for example, first look at the price and then at the reviews). Identifying this order and matching it with the browsing behavior patterns of other users can help predict which specific product in the same category the user may be most interested in, so as to provide product recommendations that better meet the user's needs. For example, on an e-commerce platform, when a user is purchasing a laptop, the user first checks the price ranges of laptops of different brands, then browses the user reviews of each product, and then pays attention to the configuration parameters of the laptops. Using algorithms such as the Markov chain model to analyze this series of browsing behaviors, it is found that users with a similar browsing order are very likely to purchase a specific laptop after checking the configuration parameters. Then, this laptop can be recommended to the user to improve the accuracy of product promotion and the probability of the user purchasing related products.

[0090] The method for identifying the user's browsing behavior pattern is as follows:

[0091] Extract each product browsing record of the user for the same type of product from the session shared information; where any one product browsing record corresponds to a specific product in the same type of product.

[0092] Count the types of relevant pages for the same type of product and assign a number to each type of relevant page; for example, the relevant pages of a certain type of product all include price pages, evaluation pages, parameter configuration pages, after-sales service pages, and so on.

[0093] Based on the product browsing records and the numbers of relevant pages, establish a browsing feature sequence and a browsing order matrix to record the user's browsing behavior pattern.

[0094] The browsing feature sequence is used to record the user's attention to each type of relevant page; the browsing feature sequence is a vector of length N; where N is the number of types of relevant pages for the same type of product; each element in the browsing feature sequence corresponds to a type of relevant page;

[0095] The method for establishing the browsing feature sequence is as follows:

[0096] Sort the N types of relevant pages and record the numbers of relevant pages based on the sorting; this sorting is arbitrary because the browsing feature sequence is only used to record the user's attention to each type of relevant page and not to record the user's browsing order between different pages.

[0097] Respectively count the total browsing duration of each type of relevant page in all product browsing records and calculate the total browsing duration of all relevant pages;

[0098] Calculate the element value corresponding to each number; form the browsing feature sequence with all element values; the element value corresponding to any number is the ratio of the total browsing duration of the relevant page corresponding to the number to the total browsing duration of all relevant pages.

[0099] The browsing order matrix is used to record the user's browsing order for different relevant pages; the method for establishing the browsing order matrix is as follows:

[0100] Sort the numbers of the N types of relevant pages horizontally and vertically to form a browsing order matrix of N rows and N columns, and initialize each element in the browsing order matrix;

[0101] Based on the product browsing records, count the user's browsing paths and the transition probabilities of each browsing path; any browsing path includes two relevant pages, namely the first page and the second page;

[0102] Let represent a browsing path where the first page is the i-th type of relevant page and the second page is the j-th type of relevant page, then the transition probability of the browsing path is denoted as , indicating the probability that the user browses the j-th type of relevant page after browsing the i-th type of relevant page; the value ranges of i and j are both 1, 2,..., N;

[0103] Assign values to the elements in the browsing order matrix based on the transition probability of each browsing path; among them, the element value at the i-th row and j-th column of the browsing order matrix is .

[0104] The described method for identifying the browsing behavior pattern of users further includes dynamically correcting the browsing order matrix, specifically including:

[0105] Identify the user's dependent browsing path; set a transition probability threshold, and mark the browsing path with a transition probability greater than the transition probability threshold as a dependent browsing path; for example, if the probability that the user browses the evaluation page after browsing the price page is higher than 70%, then mark the browsing path with the first page being the price page and the second page being the evaluation page as a dependent browsing path.

[0106] If it is detected that the user is browsing the first page of the dependent browsing path, select a page from the relevant pages not included in the dependent browsing path as the third page; push the third page to the user through a pop-up window;

[0107] Collect the operation feedback of the user on the pop-up window push; the operation feedback includes explicit positive feedback, implicit positive feedback, and negative feedback; if the user jumps and browses the third page through the pop-up window, the operation feedback is explicit positive feedback; if the user closes the pop-up window and manually jumps and browses the third page when finishing browsing the first page, the operation feedback is implicit positive feedback; if the user closes the pop-up window and does not jump to the third page when finishing browsing the first page, the operation feedback is negative feedback;

[0108] Correct the browsing order matrix based on the operation feedback. Specifically include:

[0109] Let the first page of the dependent browsing path be the p-th type of relevant page, the second page of the dependent browsing path be the q-th type of relevant page, and the third page be the r-th type of relevant page; the value ranges of p, q, and r are all 1, 2,..., N; if the user's operation feedback is explicit positive feedback, then reduce the transition probability of the browsing path of and increase the transition probability of the browsing path ; of ; is the first correction amount; if the user's operation feedback is implicit positive feedback, then reduce the transition probability of the browsing path of and increase the transition probability of the browsing path ; ; of ; ; is the preset second correction amount; and Optimality can be verified based on a large number of experiments. If the user's operation feedback is negative feedback, then keep unchanged. Preferably, if the user's operation feedback is negative feedback, then the browsing path will be added to the ignore list, indicating that when it is subsequently detected that the user is browsing the p-th relevant page, the r-th relevant page will no longer be selected as the third page for pop-up push.

[0110] Based on traditional probability statistics, the method for dynamically correcting the browsing order matrix provided in the embodiments of the present application interactively verifies the stability of the user behavior pattern through active pop-up push, and corrects the browsing order matrix in combination with the user's operation feedback on the pop-up page, so that the browsing order matrix better conforms to the user's true intention. By pushing the pop-up page, it can be verified whether the user blindly follows the historical path. If the user's browsing behavior changes significantly due to the interference of the pop-up window, it means that the original browsing order matrix fails to reflect the true needs of the user; thus, it improves the problem that the dynamic changes of the user's focus cannot be captured based on historical statistical data.

[0111] The purchase behavior pattern refers to the purchase habit that users may buy some other related products when purchasing a certain product. By using algorithms such as association rule learning to mine the implicit purchase behavior patterns in the user's product purchase records, non-intuitive product correlations can be discovered, which helps to uncover products that are not obvious but actually have a close connection, providing a basis for product investigation and promotion strategies. Taking maternal and child products as an example, after a user purchases infant milk powder, they may also purchase baby bottles, and then may purchase baby food supplements. After discovering this purchase behavior pattern through the algorithm, it is possible to accurately recommend baby bottles and food supplements to users who purchase milk powder, and also understand the potential associations between these products.

[0112] The method for identifying the user's purchase behavior pattern is as follows:

[0113] Extract the user's product purchase records from the session shared information, and extract each product purchased by the user from the product purchase records; assign a product number to any product purchased by the user;

[0114] Record the purchase time of each product purchased by the user and the corresponding product number, and establish the user's purchase time series; each position in the purchase time series corresponds to a time point (the time point is specific to a certain hour or minute of a certain year, month, and day); if the user does not purchase a product at any time point, the position corresponding to the time point is empty; if the user purchases a product at any time point, the element value of the position corresponding to the time point is the product number;

[0115] ​Set the length of the sliding window and the sliding step size; intercept the purchase time series based on the sliding window and the sliding step size; create a purchase event based on each intercepted sequence segment.

[0116] Based on the product number and the purchase event, establish a purchase event table to record the purchase behavior pattern of users.

[0117] Any column in the purchase event table corresponds to the product number of a product; any row in the purchase event table corresponds to a purchase event.

[0118] The method for establishing the purchase event table is as follows:

[0119] Create an empty purchase event table based on the number of product numbers and the number of purchase events.

[0120] Assign values to the empty purchase event table based on the product numbers included in each purchase event, specifically as follows: in any purchase event, the cell at the corresponding position in the purchase event table for any product purchased by the user is assigned a value of 1, and the cell at the corresponding position in the purchase event table for any product not purchased by the user is assigned a value of 0.

[0121] Obtain the high-intent products of the user based on the browsing behavior pattern; obtain the associated intent products of the user based on the purchase behavior pattern.

[0122] High-intent products refer to specific several products with relatively high purchase probabilities among the products of the same category browsed by the user.

[0123] The method for obtaining the high-intent products of the user based on the browsing behavior pattern is as follows:

[0124] Obtain the product browsing records and product purchase records of other users.

[0125] Based on the product browsing records and product purchase records, screen out other users who have browsed and purchased products of the same category and mark them as reference users.

[0126] Establish the browsing feature sequence and browsing order matrix for each reference user.

[0127] For any reference user, calculate the similarity between the user's browsing feature sequence and the reference user's browsing feature sequence and the similarity between the browsing order matrices and sum them to obtain the similarity between the user and the reference user.

[0128] Set a similarity threshold; mark the reference users with a similarity higher than the similarity threshold to the user as similar users.

[0129] Extract the specific models of the same type of products purchased by each similar user based on the product purchase records; count the number of purchases of each specific model of product, and mark the m products with the most purchases as the high-intent products.

[0130] Associated intent products refer to other related products that a user is likely to purchase again after purchasing a product.

[0131] The method for obtaining the associated intent products of a user based on the purchase behavior pattern is as follows:

[0132] Set the association degree threshold and the purchase frequency threshold;

[0133] Combine any two products in the purchase event table whose purchase frequencies are both greater than the purchase frequency threshold to obtain the first association group; any first association group contains two products with different product numbers;

[0134] Count the association degree of each first association group, and delete the first association groups with an association degree less than the association degree threshold; the association degree of the first association group is the number of times the first association group appears in the purchase event table divided by the total number of purchase events; for example, if a first association group contains two products with product numbers A and B respectively, and in the purchase event table, there are k rows (i.e., k purchase events) where the values of the cells corresponding to product numbers A and B are both 1, then k is the number of times the first association group appears in the purchase event table.

[0135] Merge any two first association groups and perform deduplication to obtain the second association group; count the association degree of each second association group, and delete the second association groups with an association degree less than the association degree threshold; the association degree of the second association group is the number of times the second association group appears in the purchase event table divided by the total number of purchase events; save the first association group and the second association group. The deduplication operation is because the same second association group may appear after the first association groups are merged. For example, for the three first association groups with product numbers AB, AC, and BC, after pairwise merging, 3 identical second association groups are formed, that is, the product numbers are all ABC.

[0136] Extract the list of all purchased products of the user in the most recent n days based on the product purchase records; n is a positive integer; combine the first association group and the second association group to obtain the associated intent products of the user.

[0137] The specific process of combining the first association group and the second association group to obtain the associated intent products of the user includes:

[0138] S101: If there is a second association group that contains any two purchased products, and there are products in the second association group that the user has not purchased, then mark the products that the user has not purchased in the second association group as associated intent products;

[0139] S102: If no associated product of interest is obtained in S101, obtain the associated product of interest through the first association group as follows:

[0140] If there is a first association group that contains any one of the purchased products and there are products in the first association group that the user has not purchased, mark the products in the first association group that the user has not purchased as associated products of interest;

[0141] Sort all the associated products of interest based on the association degree of the corresponding first association group, and retain the associated products of interest corresponding to the M first association groups with the largest association degrees. M is a positive integer.

[0142] Promote products to the user based on the high-interest products and the associated products of interest, including regular promotion and scenario-linked promotion; among them, the method of the regular promotion includes:

[0143] Based on the session shared information, detect the user's browsing behavior and purchase behavior in real time; when it is detected that the user has completed browsing any type of product and it is not detected that the user has purchased the same type of product, obtain the user's high-interest products and recommend high-interest products of the same type as the browsed product to the user;

[0144] When it is detected that the user has purchased a certain product, obtain the user's associated products of interest and recommend the associated products of interest that are in the same second association group or the first association group as the purchased product to the user.

[0145] The method of the scenario-linked promotion is as follows:

[0146] Obtain the user's high-interest products and the associated products of interest and form a product recommendation list;

[0147] Extract the user's device information from the session shared information; determine the common geographical location of the user's device based on the device information; combine the product purchase records and calculate the purchase probability of each product in the product recommendation list at each common geographical location; for any product A, the calculation method of its purchase probability at any common geographical location D is as follows: obtain the purchase times of all products of the same type as product A in each common geographical location in the user's product purchase record; the purchase probability of product A at the common geographical location D The calculation formula is as follows:

[0148] ;

[0149] Among them, represents the purchase times of all products of the same type as product A at the common geographical location D, represents the total purchase times of all products of the same type as product A at all common geographical locations.

[0150] Mark the common geographical location with the highest purchase probability as the associated geographical location of the corresponding product;

[0151] If it is detected that the user's device is in the associated geographical location of any product in the product recommendation list, recommend the corresponding product to the user.

[0152] Embodiment 2

[0153] This embodiment is the second embodiment of this application; based on the same inventive concept as Embodiment 1, referring to Figure 2 , this embodiment introduces a survey and promotion system based on shared information, including an information acquisition module, a behavior recognition module, a product survey module, and a product promotion module; wherein:

[0154] The information acquisition module is used to acquire the user's session shared information; including acquiring user identification, product browsing records, product purchase records, and device and environment information. These information are the basis for subsequent analysis and decision-making, and are used to identify user identity, analyze user behavior, and understand the user's device usage and purchase scenarios.

[0155] The behavior recognition module identifies the user's behavior pattern based on the session shared information; the user's behavior pattern includes browsing behavior pattern and purchase behavior pattern; this module extracts the browsing records of the user for the same type of product from the session shared information, counts the types of relevant pages and numbers them, establishes a browsing feature sequence and a browsing order matrix, so as to record the user's browsing behavior pattern and help predict the products that the user is interested in. This module is also used to acquire the user's product purchase records, extract the purchased products and number them, establish a purchase time series, create purchase events by setting a sliding window to intercept the series, and then establish a purchase event table to record the purchase behavior pattern and mine the potential associations between products.

[0156] The product survey module obtains the user's high-intent products based on the browsing behavior pattern and obtains the user's associated intent products based on the purchase behavior pattern; this module obtains the browsing and purchase records of other users, screens reference users, calculates the similarity with the target user, marks similar users, and counts the models of products purchased by similar users, so as to determine the high-intent products. The way for this module to determine the associated intent products is as follows: set the association degree threshold and the purchase times threshold, combine and count the association degree of the products in the purchase event table, merge and de-duplicate, and combine with the user's recent purchased product list to determine the associated intent products.

[0157] The product promotion module is used to promote products to users, including regular promotion and scene-linked promotion. The module performs regular promotion in the following ways: real-time detection of user browsing and purchasing behavior, recommending high-intent products when users have browsed products but have not purchased them; after users have purchased products, recommending related intention products. The module performs scene-linked promotion in the following ways: determining the common geographic locations of user devices, calculating the purchase probability of products in each common geographic location, marking the related geographic locations, and recommending the corresponding products when the user device is in the related geographic location.

[0158] The specific functional implementation of each of the above modules refers to the relevant content of the investigation and promotion method based on shared information described in Example 1, and will not be repeated here.

[0159] If the disclosed technical solution involves personal information, the product using the disclosed technical solution has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the disclosed technical solution involves sensitive personal information, the product using the disclosed technical solution has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

[0160] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0161] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose and scope of protection of the present application, all of which are within the protection of the present application.

Claims

1. A survey and promotion method based on shared information, characterized by: The following steps are involved: Get the user's session sharing information; Identify the user's behavior pattern based on the session shared information; specifically, identify the user's browsing behavior pattern and purchasing behavior pattern; Acquire the user's high-intent products based on the browsing behavior pattern; acquire the user's associated-intent products based on the purchasing behavior pattern; Product promotion is performed to users based on the high-intention products and related-intention products.

2. The survey and promotion method based on shared information as claimed in claim 1, characterized in that: The method for identifying the browsing behavior pattern of a user is as follows: Extracting each product browsing record of the user for the same type of product from the session shared information; Count the types of related pages for the same type of product and assign a number to each type of related page; Based on the product browsing records and the serial numbers of the related pages, a browsing feature sequence and a browsing order matrix are established to record the user's browsing behavior pattern.

3. The survey and promotion method based on shared information as claimed in claim 2, characterized in that: The browsing feature sequence is used to record the user's attention to each related page; The browsing feature sequence is a vector of length N, where N is the number of related pages of the same type of product; each element in the browsing feature sequence corresponds to a related page; The method to establish a browsing feature sequence is as follows: Sort the N related pages, and record the numbers of the related pages based on the sorting; Count the total browsing time of each related page in all product browsing records, and calculate the total browsing time of all related pages; Calculate the element value corresponding to each number; combine all the element values ​​into a browsing feature sequence; the element value corresponding to any number is the ratio of the total browsing time of the relevant page corresponding to the number to the total browsing time of all relevant pages.

4. The survey and promotion method based on shared information as claimed in claim 3, characterized in that: The browsing order matrix is ​​used to record the order in which users browse different related pages. The method for establishing the browsing order matrix is ​​as follows: Sort the numbers of N related pages horizontally and vertically to form a browsing order matrix of N rows and N columns, and initialize each element in the browsing order matrix; The browsing paths of users and the transition probability of each browsing path are counted based on the product browsing records; any browsing path includes two related pages, namely the first page and the second page; make Indicates that the first page is the i-th related page, and the second page is the j-th related page. The browsing path is The transition probability is denoted as , represents the probability of a user browsing the jth related page after browsing the ith related page; Based on the transition probability of each browsing path, the elements in the browsing order matrix are assigned values; where the element value of the i-th row and j-th column of the browsing order matrix is .

5. The survey and promotion method based on shared information as claimed in claim 4, characterized in that: The identifying of the browsing behavior pattern of the user also includes dynamically modifying the browsing order matrix, specifically including: Identify the dependent browsing paths of the user; set a transition probability threshold, and mark the browsing paths with a transition probability greater than the transition probability threshold as dependent browsing paths; If it is detected that the user is browsing the first page of the dependent browsing path, select a page from the related pages not included in the dependent browsing path as the third page; and push the third page to the user through a pop-up window; Collecting user feedback on pop-up window push; the feedback includes explicit positive feedback, implicit positive feedback, and negative feedback; The browsing order matrix is ​​modified based on the operational feedback.

6. The survey and promotion method based on shared information as claimed in claim 5, characterized in that: If the user jumps to and browses the third page through the pop-up window, the operation feedback is explicit positive feedback; if the user closes the pop-up window and manually jumps to and browses the third page when finishing browsing the first page, the operation feedback is implicit positive feedback; if the user closes the pop-up window and does not jump to the third page when finishing browsing the first page, the operation feedback is negative feedback; Modifying the browsing order matrix based on the operation feedback specifically includes: Let the first page that depends on the browsing path be the pth related page, the second page that depends on the browsing path be the qth related page, and the third page be the rth related page; if the user's operation feedback is explicit positive feedback, then the browsing path The transition probability Reduce , and browse to the path The transition probability Increase ; is the preset first correction amount; if the user's operation feedback is implicit positive feedback, the browsing path The transition probability Reduce , and browse to the path The transition probability Increase ; is the preset second correction value; if the user's operation feedback is negative feedback, then keep and constant.

7. The survey and promotion method based on shared information as claimed in claim 1, characterized in that: Here are some ways to identify users’ purchasing behavior patterns: Extracting the user's product purchase record from the session shared information, and extracting each product purchased by the user from the product purchase record; assigning a product number to each product purchased by the user; Record the purchase time and corresponding product number of each product purchased by the user, and establish the user's purchase time sequence; each position in the purchase time sequence corresponds to a time point; if the user does not purchase a product at any time point, the position corresponding to the time point is empty; if the user purchases a product at any time point, the element value of the position corresponding to the time point is the product number; Setting the length of the sliding window and the sliding step size; intercepting the purchase time series based on the sliding window and the sliding step size; Create a purchase event based on each sequence segment obtained by interception; A purchase event table is established based on the product number and purchase event to record the user's purchase behavior pattern.

8. The survey and promotion method based on shared information as claimed in claim 7, characterized in that: Any column in the purchase event table corresponds to a product number of a product; any row in the purchase event table corresponds to a purchase event; The method to create a purchase event table is as follows: Create an empty purchase event table based on the number of product numbers and the number of purchase events; The empty purchase event table is assigned values ​​based on the product number contained in each purchase event, as follows: in any purchase event, the cell at the corresponding position in the purchase event table for any product purchased by the user is assigned a value of 1, and the cell at the corresponding position in the purchase event table for any product not purchased by the user is assigned a value of 0.

9. The survey and promotion method based on shared information according to claim 6, characterized in that: The method of obtaining users’ high-intent products based on their browsing behavior patterns is as follows: Obtain other users' product browsing and purchase records; Based on product browsing and purchasing records, filter out other users who have browsed and purchased the same type of products and mark them as reference users; Establishing the browsing feature sequence and browsing order matrix of each reference user; For any reference user, calculate the similarity between the browsing feature sequences of the user and the reference user and the similarity between the browsing order matrices and sum them up to obtain the similarity between the user and the reference user; Setting a similarity threshold; marking reference users whose similarity with the user is higher than the similarity threshold as similar users; Based on the product purchase records, the specific models of the same type of products purchased by each similar user are extracted; the purchase times of each specific model of the product are counted, and the m products with the most purchase times are marked as the high-intention products.

10. The survey and promotion method based on shared information according to claim 8, characterized in that: The method for obtaining the user's associated intended products based on the purchase behavior pattern is as follows: Set relevance threshold and purchase number threshold; Combining any two products in the purchase event table whose purchase times are greater than the purchase times threshold to obtain a first association group; any first association group contains two products with different product numbers; Counting the relevance of each first association group, and deleting the first association group whose relevance is less than the relevance threshold; the relevance of the first association group is the number of times the first association group appears in the purchase event table divided by the total number of purchase events; Merge any two first association groups and remove duplicates to obtain a second association group; count the association degree of each second association group, and delete the second association group whose association degree is less than the association degree threshold; the association degree of the second association group is the number of times the second association group appears in the purchase event table divided by the total number of purchase events; save the first association group and the second association group; Based on the product purchase records, a list of all products purchased by the user in the last n days is extracted; n is a positive integer; and the user's associated intended products are obtained by combining the first associated group and the second associated group.

11. The survey and promotion method based on shared information according to claim 10, characterized in that: The step of combining the first association group and the second association group to obtain the user's associated intended products specifically includes: S101: If there is a second association group containing any two purchased products, and there is a product in the second association group that the user has not purchased, then mark the product in the second association group that the user has not purchased as an association intended product; S102: If no associated intended product is obtained in S101, the associated intended product is obtained through the first associated group, as follows: If the first associated group contains any purchased product, and there is a product in the first associated group that the user has not purchased, then the product in the first associated group that the user has not purchased is marked as an associated intended product; All associated intended products are sorted based on the association degrees of the corresponding first associated groups, and associated intended products corresponding to the M first associated groups with the largest association degrees are retained.

12. The survey and promotion method based on shared information according to claim 11, characterized in that: Promoting products to users based on the high-intent products and related-intent products, including conventional promotions; The conventional promotion methods include: Real-time detection of user browsing and purchasing behaviors based on session sharing information; When it is detected that the user has finished browsing any type of products and it is not detected that the user has purchased the same type of products, the user's high-intention products are obtained, and high-intention products of the same type as the browsed products are recommended to the user; When it is detected that the user has purchased a certain product, the user's associated intended products are obtained, and associated intended products in the same second associated group or first associated group as the purchased product are recommended to the user.

13. The survey and promotion method based on shared information according to claim 12, characterized in that: Product promotion to users based on the high-intent products and related-intent products also includes scene linkage promotion; the scene linkage promotion method is as follows: Obtain the user's high-intent products and related-intent products, and form a product recommendation list; Extracting the user's device information from the session shared information; determining a common geographic location of the user's device based on the device information; Based on the product purchase records, calculate the purchase probability of each product in the product recommendation list in each common geographical location; Mark the common geographical locations with the highest purchase probability as the associated geographical locations of the corresponding products; If it is detected that the user's device is in the associated geographic location of any product in the product recommendation list, the corresponding product is recommended to the user.

14. A survey and promotion system based on shared information, which is used to implement the survey and promotion method based on shared information as claimed in any one of claims 1 to 13, characterized in that: It includes information acquisition module, behavior recognition module, product investigation module and product promotion module; among which: The information acquisition module is used to obtain the user's session sharing information; The behavior recognition module recognizes the user's behavior pattern based on the session sharing information; The product research module obtains users’ high-intent products based on browsing behavior patterns, and obtains users’ related-intent products based on purchasing behavior patterns; The product promotion module is used to promote products to users, including regular promotion and scenario-linked promotion to users.

Images