Advertisement delivery method and system

Abstract

The invention provides an advertisement delivery method and system, and the method comprises the steps: obtaining user behavior data, and carrying out the preprocessing, so as to convert the user behavior data into structural data; analyzing and processing the user behavior data based on a preset construction rule to extract user interest data; extracting user historical behavior data in a preset time period from the user behavior data, and analyzing and processing the user historical behavior data to obtain a user behavior prediction result; and pushing the advertisement in the advertisement library to the user according to the user interest data and the user behavior prediction result and based on a preset pushing strategy. The interest data of the user is determined by analyzing and processing the behavior data, the behavior trend of the user is predicted, and advertisement putting is performed on the user according to the interest data and the behavior trend of the user, so that the problem that the existing advertisement putting technology cannot accurately meet the personalized requirements of the user is solved, the advertisement putting accuracy and effect are improved, and the user experience is improved. And the user experience is improved.

Description

Technical Field

[0001] The present application relates to the technical field of advertisement delivery, and in particular to an advertisement delivery method and system. Background Art

[0002] With the rapid development of Internet technology, the methods and channels of advertising are becoming increasingly diversified.

[0003] Most existing advertising technologies are based on broad user group characteristics and cannot accurately meet the personalized needs of each user. For example, advertising is only pushed based on basic information such as age and gender of users, which makes it difficult to accurately capture users' interests and real-time behavior changes, resulting in a waste of advertising resources, poor advertising effects, and reduced user experience. Summary of the invention

[0004] In view of this, this application proposes an advertisement delivery method and system, the specific scheme is as follows:

[0005] In the first part, the present application proposes an advertisement delivery method, which is applied to a system including an advertisement library. The method comprises:

[0006] Acquire user behavior data from one or more preset data sources, and pre-process the user behavior data to convert the user behavior data into structured data;

[0007] The user behavior data converted into structured data is analyzed and processed based on preset construction rules by using a preset user interest construction model to extract user interest data;

[0008] Extracting user historical behavior data within a preset time period from the user behavior data converted into structured data, and analyzing and processing the user historical behavior data through a preset prediction model to obtain a user behavior prediction result;

[0009] The advertisements in the advertisement library are pushed to the user according to the user interest data and the user behavior prediction result and based on a preset push strategy.

[0010] In a specific embodiment, the user behavior data includes: browsing record data, and / or search record data, and / or purchase record data, and / or social media interaction data.

[0011] In a specific embodiment, the “preprocessing the user behavior data to convert the user behavior data into structured data” includes:

[0012] Identify the user behavior data and generate text data;

[0013] The text data is uniformly formatted and converted based on a preset recognizable format to generate structured data.

[0014] In a specific embodiment, it is characterized in that the “using a preset user interest construction model to analyze and process the user behavior data converted into structured data based on preset construction rules to extract user interest data” includes:

[0015] After performing semantic analysis on the user behavior data converted into structured data using natural language processing technology and machine learning algorithms, keywords are extracted from the user behavior data converted into structured data based on preset rules;

[0016] According to the occurrence frequency and time period of the user behavior corresponding to each keyword, a weight is assigned to each keyword to form user interest data.

[0017] In a specific embodiment, the “analyzing and processing the user historical behavior data by a preset prediction model to obtain a user behavior prediction result” specifically includes:

[0018] The user historical behavior data is analyzed and processed based on time series analysis and deep learning algorithms to obtain user behavior prediction results.

[0019] In a specific embodiment, the “pushing the advertisement in the advertisement library to the user according to the user interest data and the user behavior prediction result and based on a preset push strategy” specifically includes:

[0020] Calculating the similarity between the advertisement features in the advertisement library and the user interest data and the user behavior prediction result;

[0021] Based on a preset push strategy, the advertisement corresponding to the similarity reaching a preset value is pushed to the corresponding user.

[0022] In a specific embodiment, it also includes:

[0023] One or any multiple of the user behavior data, the user interest data, and the user behavior prediction results are associated with the user and then stored.

[0024] In the second part, the present application proposes an advertisement delivery system, which is applied to a system including an advertisement library, and includes:

[0025] An acquisition module, used to acquire user behavior data from one or more preset data sources, and pre-process the user behavior data to convert the user behavior data into structured data;

[0026] An extraction module, used to analyze and process the user behavior data converted into structured data based on preset construction rules through a preset user interest construction model to extract user interest data;

[0027] A processing module, used to extract user historical behavior data within a preset time period from the user behavior data converted into structured data, and analyze and process the user historical behavior data through a preset prediction model to obtain a user behavior prediction result;

[0028] A push module is used to push the advertisements in the advertisement library to the user according to the user interest data and the user behavior prediction result and based on a preset push strategy.

[0029] In a specific embodiment, the “preprocessing the user behavior data to convert the user behavior data into structured data” includes:

[0030] An identification module, used to identify the user behavior data and generate text data;

[0031] The conversion module is used to perform a unified format conversion on the text data based on a preset recognizable format to generate structured data.

[0032] In a specific embodiment, the “analyzing and processing the user behavior data converted into structured data based on preset construction rules by using a preset user interest construction model to extract user interest data” includes:

[0033] An analysis module, configured to perform semantic analysis on the user behavior data converted into structured data using natural language processing technology and machine learning algorithms, and extract keywords from the user behavior data converted into structured data based on preset rules;

[0034] The weighting module is used to assign a weight to each of the keywords according to the occurrence frequency and time period of the user behavior corresponding to each of the keywords, so as to form user interest data.

[0035] Beneficial effects:

[0036] This application obtains user behavior data, analyzes and processes the behavior data to determine the user's interest data, and predicts the user's behavior trend. It delivers advertisements to users based on the user's interest data and behavior trend, thereby solving the problem that existing advertising delivery technology cannot accurately meet the personalized needs of users, improving the accuracy and effect of advertising delivery, and improving user experience. It optimizes the configuration of advertising resources, avoids the waste of advertising resources, and improves the efficiency of advertising delivery. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a flow chart of an advertisement delivery method according to an embodiment of the present application;

[0038] Figure 2 is a schematic diagram of an advertisement delivery system module according to an embodiment of the present application;

[0039] Figure 3 This is a diagram of the architecture of the advertising delivery method according to an embodiment of the present application;

[0040] Figure 4 This is a flowchart of pushing advertisements according to an embodiment of the present application.

[0041] Figure numerals: 1 - acquisition module; 2 - extraction module; 3 - processing module; 4 - push module. DETAILED DESCRIPTION

[0042] Hereinafter, various embodiments disclosed in the present application will be described more fully. The present application may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments disclosed in the present application to the specific embodiments disclosed herein, but the present application should be understood to cover all adjustments, equivalents and / or alternatives falling within the spirit and scope of the various embodiments disclosed in the present application.

[0043] The terms used in the various embodiments disclosed in the present application are only used to describe the purpose of specific embodiments and are not intended to limit the various embodiments disclosed in the present application. As used herein, the singular form is intended to also include the plural form, unless the context clearly indicates otherwise. Unless otherwise specified, all terms used here (including technical terms and scientific terms) have the same meaning as the meaning generally understood by ordinary technicians in the field of various embodiments disclosed in the present application. The terms (such as the terms defined in the dictionary generally used) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized meaning or an overly formal meaning, unless clearly defined in the various embodiments disclosed in the present application.

[0044] Example 1

[0045] Embodiment 1 of the present application discloses an advertising delivery method. This embodiment obtains user behavior data, analyzes and processes the behavior data to determine the user's interest data, and predicts the user's behavior trend. It delivers advertisements to the user based on the user's interest data and behavior trend, thereby solving the problem that existing advertising delivery technology cannot accurately meet the personalized needs of users, improving the accuracy and effect of advertising delivery, and improving user experience. Optimize advertising resource allocation, avoid waste of advertising resources, and improve the efficiency of advertising delivery. The specific process is as attached in the specification. Figure 1 , Attachment Figure 3 and attached Figure 4The specific plan is as follows:

[0046] An advertisement delivery method is applied to a system including an advertisement library, as shown in the attached manual. Figure 1 As shown, the method includes:

[0047] 101. Obtain user behavior data from one or more preset data sources, and pre-process the user behavior data to convert the user behavior data into structured data. In practical applications, user behavior data on multiple platforms can be collected, including but not limited to browsing records, search records, purchase records, social interaction records, etc. The collected user behavior data is cleaned and pre-processed to remove invalid data and noise.

[0048] 102. The user behavior data converted into structured data is analyzed and processed based on preset construction rules through the preset user interest construction model to extract user interest data. In practical applications, natural language processing technology and machine learning algorithms are used to analyze the acquired user behavior data to extract user interest data and present it in the form of keywords or other preset forms. This step aims to accurately identify and extract key data that reflects user preferences, needs and interests, and provide strong data support for subsequent user portrait drawing, personalized service recommendation and market strategy formulation.

[0049] 103. Extract the user's historical behavior data within a preset time period from the user behavior data converted into structured data, and analyze and process the user's historical behavior data through a preset prediction model to obtain the user behavior prediction result. Extract the user's historical behavior data within a preset time period from the user behavior data. These data cover various interactions and operation records of the user within a specific time range. In practical applications, the preset time period can be adjusted according to actual needs. And use the pre-built and trained prediction model to conduct in-depth analysis and processing of these extracted user historical behavior data. The prediction model is based on time series analysis and deep learning algorithms to model the user's historical behavior data and predict the user's future behavior trends. For example, by analyzing the user's purchase behavior cycle in the past few months, predict the time and product category that the user may purchase next time. Through this analysis process, the system can obtain the prediction results of user behavior, better understand the user's future needs and behavior trends, and make more accurate and effective decisions.

[0050] 104. Push ads from the ad library to users based on user interest data and user behavior prediction results and based on preset push strategies. Based on user interest data and user behavior prediction results, the system will comprehensively consider the user's preferences, historical behaviors, and possible future trends, and then select ads from the ad library based on preset push strategies. These push strategies may include multiple dimensions such as personalized recommendations, interest matching rankings, time sensitivity considerations, and ad frequency control. Through such comprehensive analysis and strategy application, the system can ensure that the ads pushed to users are in line with their interests and expected behaviors, thereby increasing the attractiveness and conversion rate of ads, while also improving user experience and satisfaction.

[0051] In a specific embodiment, user behavior data includes: browsing record data, and / or search record data, and / or purchase record data, and / or social media interaction data. In this embodiment, user behavior data covers multiple aspects, including but not limited to browsing record data, which records the pages, products or services that users have browsed on websites or applications; search record data, which reflects the information or goods that users have found through search engines or on-site search functions; purchase record data, which records the user's purchase history in detail, including key information such as purchase time, product category, price, etc.; and social media interaction data, which covers users' interactive behaviors such as likes, comments, and sharing on social media platforms. These data together constitute a comprehensive view of user behavior data, providing rich materials for in-depth analysis of user preferences, interests, and behavior patterns.

[0052] In a specific embodiment, “preprocessing the user behavior data to convert the user behavior data into structured data” includes:

[0053] Identify user behavior data and generate text data;

[0054] The text data is uniformly formatted and converted based on a preset recognizable format to generate structured data.

[0055] In this embodiment, first, the acquired user behavior data is fully identified, and the unstructured raw data is converted into text data through parsing and sorting, and then the text data is uniformly formatted and converted according to a pre-set standard format that can be recognized and understood by the machine. This step ensures the standardization and consistency of the data, and ultimately generates structured data with a clear structure that is easy to analyze and apply, laying a solid foundation for subsequent data mining, analysis and application.

[0056] In a specific embodiment, “using a preset user interest construction model to analyze and process the user behavior data converted into structured data based on preset construction rules to extract user interest data” includes:

[0057] After using natural language processing technology and machine learning algorithms to perform semantic analysis on the user behavior data converted into structured data, keywords are extracted from the user behavior data converted into structured data based on preset rules;

[0058] According to the frequency and time period of user behavior corresponding to each keyword, each keyword is assigned a weight to form user interest data.

[0059] In this embodiment, advanced natural language processing technology and machine learning algorithms are first used to conduct in-depth semantic analysis on the user behavior data that has been converted into structured data, and keywords that can summarize the user's interest orientation are accurately extracted from these data.

[0060] Furthermore, according to the frequency and time period of user behaviors corresponding to each keyword, by analyzing the frequency density and time distribution characteristics of these behavior data, a weight value is scientifically and reasonably assigned to each keyword to build a user interest model. For example, if a user has recently frequently searched and browsed content about outdoor sports, the keyword "outdoor sports" will have a higher weight in the user interest model. This step not only takes into account the direct frequency of behavior, but also combines the time factor to reflect the persistence and recency of user interests, thereby forming user interest data.

[0061] In a specific embodiment, “analyzing and processing the user's historical behavior data through a preset prediction model to obtain a user behavior prediction result” specifically includes:

[0062] Based on time series analysis and deep learning algorithms, user historical behavior data is analyzed and processed to obtain user behavior prediction results.

[0063] In this embodiment, time series analysis technology is first used to carefully examine the time dimension in the user's historical behavior data to identify the trend of user behavior changes over time, periodic laws, and possible anomalies. This step helps to understand the temporal dynamics of user behavior and provides a temporal reference framework for subsequent predictions. Further, combined with deep learning algorithms, such as recurrent neural networks (RNN) or their variants (such as long short-term memory networks LSTM, gated recurrent units GRU, etc.), the user's historical behavior data is deeply mined. These algorithms can automatically learn high-dimensional features and nonlinear relationships in user behavior data, thereby more accurately predicting the user's future behavior trends. This provides a scientific basis and decision-making support for market strategy formulation, personalized recommendation system optimization, etc.

[0064] In a specific embodiment, “pushing advertisements in the advertisement library to users based on user interest data and user behavior prediction results and based on a preset push strategy” specifically includes:

[0065] Calculate the similarity between the ad features in the ad library and the user interest data and user behavior prediction results;

[0066] Based on the preset push strategy, the advertisement corresponding to the similarity reaching the preset value is pushed to the corresponding user.

[0067] In this embodiment, the system first conducts a comprehensive analysis of the ads in the ad library and extracts their key features, which may include the ad's theme, keywords, product category, target audience positioning, etc. Then, using advanced matching algorithms, it calculates the similarity between these ad features and user interest data, as well as the correlation between the ad and the user behavior prediction results. This step is intended to identify the ad content that best matches the user's current interests and possible future behaviors, ensuring the relevance and accuracy of the push.

[0068] Furthermore, based on the preset push strategy, the system will further filter and sort these ads. Push strategies may involve multiple dimensions, such as the relevance threshold between ads and user interests, the novelty of ads, the frequency limit of ads, the historical performance of ads, etc. Based on these strategy settings, the system will filter out ads whose similarity reaches or exceeds the preset value (i.e., the threshold that meets the push conditions), and accurately deliver them to the corresponding users through appropriate channels and methods, such as in-app notifications, emails, social media pushes, etc. Not only does it take into account the matching degree between ads and users, but it also focuses on optimizing the timing and method of push, ensuring the accuracy and effectiveness of ad push and improving the user experience.

[0069] In a specific embodiment, it also includes:

[0070] One or more of the user behavior data, user interest data, and user behavior prediction results are associated with the user and stored. By binding different types of data to the user, the system can build a comprehensive user portrait, which not only reflects the user's past behavior (through user behavior data), but also reveals the user's current interests (through user interest data) and predictions of future behavior (through user behavior prediction results). This provides a solid foundation for subsequent personalized services, precision marketing, and user behavior analysis.

[0071] At the same time, the data storage process also emphasizes security and privacy protection. All user-related data will be encrypted and stored in a secure environment that complies with industry standards and laws and regulations. Only authorized systems or personnel can access this data, ensuring the confidentiality and security of user information.

[0072] Example 2

[0073] Embodiment 2 of the present application discloses an advertisement delivery system, which systematizes an advertisement delivery method of embodiment 1 to make it more practical. The overall structure diagram of the advertisement delivery system is shown in the attached manual. Figure 2 The specific plan is as follows:

[0074] This embodiment provides an advertisement delivery system, which is applied to a system including an advertisement library. The system includes:

[0075] An acquisition module 1 is used to acquire user behavior data from one or more preset data sources, and pre-process the user behavior data to convert the user behavior data into structured data;

[0076] Extraction module 2, used to analyze and process the user behavior data converted into structured data based on preset construction rules through a preset user interest construction model to extract user interest data;

[0077] Processing module 3, used to extract user historical behavior data within a preset time period from the user behavior data converted into structured data, and analyze and process the user historical behavior data through a preset prediction model to obtain a user behavior prediction result;

[0078] The push module 4 is used to push advertisements in the advertisement library to users according to the user interest data and the user behavior prediction results and based on a preset push strategy.

[0079] For example, in an e-commerce platform, the acquisition module collects data such as user browsing records, purchase history, and search keywords in real time. The extraction module performs semantic analysis on the user's search keywords to extract the product categories and brands that the user is interested in. For example, if the user often searches for "smart watches", the categories and brands related to "smart watches" are included in the user interest model and given a higher weight according to the search frequency. Furthermore, the processing module analyzes the user's purchase cycle and browsing pattern to predict the time when the user may buy a smart watch again. Assume that based on historical data, it is predicted that the user has a high probability of buying a smart watch in the next month. Finally, the push module selects smart watch-related advertisements from the advertising library, including smart watch advertisements of different brands, and sorts them according to the similarity between the advertising content and the user's interests. And when the user logs in to the e-commerce platform to browse the relevant product page next time, the smart watch advertisement with a high matching degree is displayed in a prominent position on the page. If it is predicted that the user's purchase time is approaching, the user can also be reminded to pay attention to the relevant advertisements by push notification.

[0080] For another example, in a social media platform, the acquisition module collects user's published content, likes, comments, and attention behavior data. The extraction module analyzes the topics and keywords of the user's published content, as well as the types of accounts that the user follows, to build a user interest model. For example, if a user often publishes content about travel and follows travel bloggers, then "travel" will become an important tag of the user's interest. Further, the processing module predicts the time when the user is most likely to browse social media based on the user's activity level and social interaction mode in different time periods. For example, users are usually more active between 8 and 10 pm. Finally, the push module selects travel-related advertisements for the "travel" interest tag, such as tourist attraction recommendations, travel agency advertisements, and travel equipment advertisements, and further optimizes the advertisement matching based on the user's geographic location and historical travel behavior (if any). For example, if the user is located in a coastal city and has a record of coastal travel, seaside vacation-related advertisements may be pushed first. In the user's predicted active time period, matching travel advertisements are inserted into the user's information flow. At the same time, according to the user's interactive habits, different advertising forms are adopted, such as picture ads, video ads, or interactive ads, to improve user participation and advertising effectiveness.

[0081] In a specific embodiment, “preprocessing the user behavior data to convert the user behavior data into structured data” includes:

[0082] The recognition module is used to recognize user behavior data and generate text data;

[0083] The conversion module is used to perform unified formatting conversion on text data based on a preset recognizable format to generate structured data.

[0084] In a specific embodiment, “using a preset user interest construction model to analyze and process the user behavior data converted into structured data based on preset construction rules to extract user interest data” includes:

[0085] An analysis module is used to perform semantic analysis on the user behavior data converted into structured data using natural language processing technology and machine learning algorithms, and to extract keywords from the user behavior data converted into structured data based on preset rules;

[0086] The weighting module is used to assign weights to each keyword according to the frequency and time period of the user behavior corresponding to each keyword, so as to form user interest data.

[0087] In a specific embodiment, the user behavior data includes: browsing record data, and / or search record data, and / or purchase record data, and / or social media interaction data.

[0088] In a specific embodiment, “analyzing and processing the user's historical behavior data through a preset prediction model to obtain a user behavior prediction result” specifically includes:

[0089] Based on time series analysis and deep learning algorithms, user historical behavior data is analyzed and processed to obtain user behavior prediction results.

[0090] In a specific embodiment, “pushing advertisements in the advertisement library to users based on user interest data and user behavior prediction results and based on a preset push strategy” specifically includes:

[0091] Calculate the similarity between the ad features in the ad library and the user interest data and user behavior prediction results;

[0092] Based on the preset push strategy, the advertisement corresponding to the similarity reaching the preset value is pushed to the corresponding user.

[0093] In a specific embodiment, it also includes:

[0094] One or more of the user behavior data, user interest data, and user behavior prediction results are associated with the user and stored.

[0095] This embodiment proposes an advertisement delivery system, which systematizes the advertisement delivery method of Embodiment 1 to make it more practical.

[0096] This application obtains user behavior data, analyzes and processes the behavior data to determine the user's interest data, and predicts the user's behavior trend. It delivers advertisements to users based on the user's interest data and behavior trend, thereby solving the problem that existing advertising delivery technology cannot accurately meet the personalized needs of users, improving the accuracy and effect of advertising delivery, and improving user experience. It optimizes the configuration of advertising resources, avoids the waste of advertising resources, and improves the efficiency of advertising delivery.

[0097] Those skilled in the art will appreciate that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily required for implementing the present invention. Those skilled in the art will appreciate that the modules in the devices in the implementation scenario may be distributed in the devices in the implementation scenario according to the description of the implementation scenario, or may be changed accordingly and located in one or more devices different from the present implementation scenario.

Claims

1. An advertisement delivery method, characterized in that: Applied in a system including an advertisement library, the method comprises: Acquire user behavior data from one or more preset data sources, and pre-process the user behavior data to convert the user behavior data into structured data; The user behavior data converted into structured data is analyzed and processed based on preset construction rules by using a preset user interest construction model to extract user interest data; Extracting user historical behavior data within a preset time period from the user behavior data converted into structured data, and analyzing and processing the user historical behavior data through a preset prediction model to obtain a user behavior prediction result; The advertisements in the advertisement library are pushed to the user according to the user interest data and the user behavior prediction result and based on a preset push strategy.

2. The method according to claim 1, characterized in that The user behavior data includes: browsing record data, and / or search record data, and / or purchase record data, and / or social media interaction data.

3. The method according to claim 1, characterized in that The “preprocessing the user behavior data to convert the user behavior data into structured data” includes: Identify the user behavior data and generate text data; The text data is uniformly formatted and converted based on a preset recognizable format to generate structured data.

4. The method according to claim 1, characterized in that: The “analyzing and processing the user behavior data converted into structured data based on preset construction rules by using a preset user interest construction model to extract user interest data” includes: After performing semantic analysis on the user behavior data converted into structured data using natural language processing technology and machine learning algorithms, keywords are extracted from the user behavior data converted into structured data based on preset rules; According to the occurrence frequency and time period of the user behavior corresponding to each keyword, a weight is assigned to each keyword to form user interest data.

5. The method according to claim 1, characterized in that The “analyzing and processing the user historical behavior data by a preset prediction model to obtain a user behavior prediction result” specifically includes: The user historical behavior data is analyzed and processed based on time series analysis and deep learning algorithms to obtain user behavior prediction results.

6. The method according to claim 1, characterized in that The “pushing the advertisements in the advertisement library to the user according to the user interest data and the user behavior prediction result and based on a preset push strategy” specifically includes: Calculating the similarity between the advertisement features in the advertisement library and the user interest data and the user behavior prediction result; Based on a preset push strategy, the advertisement corresponding to the similarity reaching a preset value is pushed to the corresponding user.

7. The method according to claim 1, characterized in that Also includes: One or any multiple of the user behavior data, the user interest data, and the user behavior prediction results are associated with the user and then stored.

8. An advertisement delivery system, characterized in that: Used to implement the advertisement delivery method according to any one of claims 1 to 7, the system comprises: An acquisition module, used to acquire user behavior data from one or more preset data sources, and pre-process the user behavior data to convert the user behavior data into structured data; An extraction module, used to analyze and process the user behavior data converted into structured data based on preset construction rules through a preset user interest construction model to extract user interest data; A processing module, used to extract user historical behavior data within a preset time period from the user behavior data converted into structured data, and analyze and process the user historical behavior data through a preset prediction model to obtain a user behavior prediction result; A push module is used to push the advertisements in the advertisement library to the user according to the user interest data and the user behavior prediction result and based on a preset push strategy.

9. The system according to claim 8, characterized in that The “preprocessing the user behavior data to convert the user behavior data into structured data” includes: An identification module, used to identify the user behavior data and generate text data; The conversion module is used to perform a unified format conversion on the text data based on a preset recognizable format to generate structured data.

10. The system according to claim 8, characterized in that The “analyzing and processing the user behavior data converted into structured data based on preset construction rules by using a preset user interest construction model to extract user interest data” includes: An analysis module, configured to perform semantic analysis on the user behavior data converted into structured data using natural language processing technology and machine learning algorithms, and extract keywords from the user behavior data converted into structured data based on preset rules; The weighting module is used to assign a weight to each of the keywords according to the occurrence frequency and time period of the user behavior corresponding to each of the keywords, so as to form user interest data.

Images