Implementation method and system for precisely putting home page recommended content based on user portrait

Abstract

The invention discloses an implementation method and system for precisely putting home page recommended content based on a user portrait, and the method comprises the steps: enabling a system to comprehensively record various behavior data of a user in a process that the user uses an application; cleaning and de-noising the collected original data, and storing the data in a database; the background service generates a plurality of behavior tags for each user by analyzing the behavior data of the user and using a predefined algorithm and rule; the background configures specific display content for each user label and rule; and when the user enters, displaying specific browsing contents of the home page according to the divided rules and labels. Based on the detailed and dynamically updated user portrait, highly personalized content recommendation can be realized, and the relevance of the recommended content and the user satisfaction are remarkably improved. The personalized recommendation not only improves the browsing experience of the user, but also increases the residence time and the use frequency of the platform, meanwhile, the advertisement putting is more accurate, and the click rate and the conversion rate of the advertisement are effectively improved.

Description

Technical Field

[0001] The present invention relates to the fields of information collection, data processing and statistics, and in particular to a method and system for accurate content delivery based on user portraits. Background Art

[0002] Most existing homepage recommendation systems rely on general recommendation algorithms: recommending content based on users' historical behaviors and the preferences of similar users, which fails to fully utilize users' personalized data, resulting in low relevance of recommended content.

[0003] Traditional methods do not have a deep enough understanding of user needs and cannot capture changes in user interests in real time, resulting in poor user experience. In order to improve the accuracy of recommended content, a personalized recommendation method based on user portraits is needed, which can analyze user behavior in real time and dynamically adjust recommended content based on user portraits. Summary of the invention

[0004] In order to solve the problem that the recommended content in the existing homepage recommendation system has low relevance and cannot fully utilize the personalized data of users, the present invention provides a method for implementing the precise delivery of recommended content on the homepage based on user portraits. The method comprehensively collects and analyzes user behavior data, builds accurate user portraits, and generates and adjusts personalized recommended content in real time based on user portraits, thereby significantly improving the relevance of recommended content and user satisfaction. The method realizes personalized content recommendation and precise advertising delivery by analyzing user behavior data and building user portraits, thereby improving user experience and marketing effects.

[0005] To achieve the above purpose, the technical solution adopted by the present invention is: a method for accurately delivering homepage recommended content based on user portraits, including the following steps:

[0006] S1. The system comprehensively records various user behavior data during the user's use of the application, including but not limited to page browsing history, search history, click history, purchase history, collection history and social interaction;

[0007] S2. Clean and denoise the collected raw behavioral data, filter out invalid and erroneous data, convert the format, standardize the data, ensure the accuracy of subsequent analysis, and store it in the database;

[0008] S3, the backend service generates multiple behavior tags for each user by analyzing the user's behavior data and using predefined algorithms and rules; these tags describe the user's interests, preferences, and behavior patterns;

[0009] S4. The backend system configures content based on different regional sections of the homepage and multiple dimensions:

[0010] User tags match personalized content based on user interests, behaviors, and attributes;

[0011] The display duration distinguishes between permanent content that is displayed for a long time and temporary content that is displayed for a short time;

[0012] Scheduled display sets the content to be displayed during a specific time period;

[0013] The access platform adjusts the content and format of the display according to the device used by the user, and the displayed content includes news articles, product information, advertisements, videos, and social media content;

[0014] When users use the application, specific browsing content on the homepage is displayed according to the divided rules and tags.

[0015] Optionally, in some embodiments of the present invention, step S1 includes: data collection includes defining the names of events to be collected, including page browsing, searching, clicking, purchasing, collecting, and social interaction, and defining the information content to be collected for each event, including page URL, search keywords, click button ID, purchased product ID, collection time, and social interaction type; when users trigger these events, the client records relevant information and extracts required data, encrypts the data using the RSA algorithm, and stores the encrypted data in the Redis cache for subsequent upload.

[0016] Optionally, in some embodiments of the present invention, step S2 includes:

[0017] User behavior data is consumed by Redis queues regularly to filter out invalid and erroneous data; data of different event types are classified and processed and stored in MongoDB for subsequent analysis and processing.

[0018] Optionally, in some embodiments of the present invention, step S3 includes:

[0019] The backend service analyzes user behavior data and uses predefined algorithms and rules to generate multiple behavior tags for each user. These tags describe the user's interests, preferences, and behavior patterns. It uses machine learning algorithms to continuously optimize and update behavior tags to ensure the accuracy and real-time nature of recommended content.

[0020] Optionally, in some embodiments of the present invention, the system generates tags using rule-based algorithms, statistical algorithms, or behavioral feature algorithms based on recorded user behavior data.

[0021] Optionally, in some embodiments of the present invention, step S4 includes:

[0022] The backend configures specific display content for each user tag and rule, including news articles, product information, advertisements, videos, and social media content. When a user comes in, specific browsing content on the homepage is displayed according to the divided rules and tags. The system supports A / B testing and multiple recommendation strategies to evaluate and optimize recommendation effects and enhance user experience.

[0023] On the other hand, the present invention provides a system for implementing accurate delivery of homepage recommended content based on user portraits, comprising:

[0024] User behavior data collection module, which comprehensively records various user behavior data, including but not limited to page browsing history, search history, click history, purchase history, collection history and social interaction;

[0025] The data preprocessing module is used to clean and denoise the collected raw data. The redis queue is used for consumption. The system obtains data from the queue, filters out invalid and erroneous data, and stores each event in the mongodb database for subsequent in-depth analysis and processing. Data indexes are established in mongodb to improve the efficiency of data retrieval and analysis, and support complex query operations.

[0026] User tag building module, which analyzes user behavior data to build user tags. Based on the recorded user behavior data, the system uses predefined algorithms and rules to generate tags;

[0027] The homepage content construction module configures content according to different areas of the homepage and combines multiple dimensions: user tags match personalized content according to user interests, behaviors and attributes; display duration distinguishes between long-term permanent content and short-term temporary content; scheduled display sets content to be displayed during specific time periods; the access platform adjusts the display content and format according to the device used by the user; when the user opens the application, the system dynamically generates and displays specific browsing content on the homepage based on pre-set rules and user tags.

[0028] The beneficial effects of the present invention are:

[0029] (1) By building a detailed and dynamically updated user profile, the system can accurately identify the user's interests and needs, thereby achieving highly personalized content recommendations. The system generates precise behavioral tags by analyzing the user's browsing history, search history, purchase history, and click behavior. These tags help the system better understand the user's preferences and habits, thereby greatly improving the relevance of recommended content and user satisfaction.

[0030] (2) By using user behavior tags and advanced recommendation algorithms, the system can provide each user with customized homepage content. This content includes news articles, product information, advertisements, videos, social media content, etc., which greatly improves the user's browsing experience. Personalized recommendations not only make it easier for users to find content that interests them, but also increase the time users stay on the platform and the frequency of use.

[0031] (3) The system can dynamically adjust recommended content based on the user's real-time behavior data to ensure that the recommended content always matches the user's current interests and needs. No matter how the user's interests change, the system can respond quickly and provide the most timely and relevant recommended content. This real-time nature makes the recommendation system more flexible and intelligent, and the user experience more smooth and satisfying.

[0032] (4) Through accurate user portraits and behavior analysis, the system can deliver advertisements more effectively. By using users’ interest tags and behavior patterns, the system can accurately push advertisements to users who are most likely to be interested in them, thereby significantly improving the click-through rate and conversion rate of advertisements. This kind of precision marketing not only improves the effectiveness of advertisements, but also increases advertiser satisfaction and optimizes the overall marketing strategy.

[0033] (5) Through personalized content recommendations, the system can significantly increase the user's stay time and interaction rate on the platform. Users can not only find the content they are interested in more quickly, but also enjoy more personalized services and experiences, which greatly increases the user's dependence on and loyalty to the platform. High levels of user stickiness and engagement not only help the platform maintain its activity, but also bring more user traffic and business opportunities. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flowchart of the implementation method of the present invention for accurately delivering recommended content on the homepage based on user portraits. DETAILED DESCRIPTION

[0035] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0036] The present invention discloses a method and system for implementing accurate delivery of home page recommended content based on user portraits, which is an implementation method for accurately delivering home page recommended content based on user portraits, comprising:

[0037] The system comprehensively records various user behavior data while the user is using the application;

[0038] Clean and denoise the collected raw data and store them in the database;

[0039] The backend service generates multiple behavior tags for each user by analyzing the user's behavior data and using predefined algorithms and rules;

[0040] The background configures specific display content for each user tag and rule;

[0041] When a user comes in, specific browsing content on the homepage is displayed according to the divided rules and tags.

[0042] Based on detailed and dynamically updated user portraits, the present invention can achieve highly personalized content recommendations, significantly improving the relevance of recommended content and user satisfaction. Personalized recommendations not only improve the user's browsing experience, but also increase the platform's dwell time and usage frequency, while making advertising delivery more accurate, effectively improving the click-through rate and conversion rate of advertising.

[0043] Specifically, in the present invention, the method for implementing the accurate delivery of homepage recommended content based on user portraits includes the following steps:

[0044] S1. When users use the application, the system comprehensively records various user behavior data, including but not limited to page browsing history, search history, click history, purchase history, collection history and social interaction.

[0045] S2. Clean and denoise the collected raw data, filter out invalid and erroneous data, convert the format, standardize the data, ensure the accuracy of subsequent analysis, and store it in the database.

[0046] S3. The backend service generates multiple behavior tags for each user using predefined algorithms and rules by analyzing the user's behavior data (e.g., browsing history, search history, click behavior, etc.). These tags describe the user's interests, preferences, and behavior patterns. Based on the recorded user behavior data, the system can use several methods to generate tags: rule-based algorithms, such as marking as "channel users" based on the user's channel source, or marking as "high-value users" based on certain preset conditions; based on statistical algorithms, such as counting the content categories that users browse most frequently and marking them as corresponding "interest tags"; based on behavioral feature algorithms, treating the user's behavior as a time series and analyzing the temporal pattern of the user's behavior, which can discover the user's habits, interest change trends, etc., and mark them with corresponding tags.

[0047] S4. The backend system configures content according to different regions of the homepage and multiple dimensions: user tags match personalized content based on user interests, behaviors, attributes and other characteristics; display duration distinguishes between long-term resident content and short-term temporary content; scheduled display sets content to be displayed during specific time periods, such as morning news and lunchtime recommendations; the access platform adjusts the display content and format based on the device used by the user (such as mobile phones, tablets, PCs, etc.). Display content includes news articles, product information, advertisements, videos, and social media content; when users use the application, specific browsing content on the homepage is displayed according to the divided rules and tags.

[0048] In a further technical solution, the comprehensive recording of various user behavior data in step S1 includes:

[0049] Record the pages visited by users and save the related button behaviors clicked by users to the redis cache.

[0050] In this technical solution, it is necessary to define the collection event name and the information that needs to be reported, such as: including but not limited to page browsing history, search history, click history, purchase history, collection history and social interaction, and perform RSA encryption on the information and save the encrypted information in the redis cache.

[0051] In a further technical solution, step S2 includes:

[0052] Regularly consume user behavior data in the redis queue, filter out invalid and erroneous data, and store each event in mongodb.

[0053] In this technical solution, because there is too much user behavior data, redis queues are used for consumption. The system obtains data from the queue, filters out invalid and erroneous data, and stores each event in the mongodb database for subsequent in-depth analysis and processing. Data indexes are established in mongodb to improve the efficiency of data retrieval and analysis and support complex query operations.

[0054] In a further technical solution, step S3 includes:

[0055] The backend service generates multiple behavior tags for each user by analyzing the user's behavior data (e.g., browsing history, search history, click behavior, etc.) using predefined algorithms and rules. These tags describe the user's interests, preferences, and behavior patterns.

[0056] In a further technical solution, step S4 includes:

[0057] The backend configures specific display content for each user tag and rule. These contents can include news articles, product information, advertisements, videos, social media content, etc. When a user comes in, the homepage specific browsing content is displayed according to the divided rules and tags.

[0058] In this technical solution, the home page content and display rules are configured for the user's various tags in the background. When the user visits the home page, the filtered and sorted recommended tag configuration content is displayed to the user in a user-friendly manner.

[0059] In addition, the present invention also provides a system for implementing the precise delivery of homepage recommended content based on user portraits, including:

[0060] User behavior data collection module, which comprehensively records various user behavior data, including but not limited to page browsing history, search history, click history, purchase history, collection history and social interaction;

[0061] The data preprocessing module is used to clean and denoise the collected raw data. The redis queue is used for consumption. The system obtains data from the queue, filters out invalid and erroneous data, and stores each event in the mongodb database for subsequent in-depth analysis and processing. Data indexes are established in mongodb to improve the efficiency of data retrieval and analysis, and support complex query operations.

[0062] User tag building module, which analyzes user behavior data to build user tags. Based on the recorded user behavior data, the system uses predefined algorithms and rules to generate tags;

[0063] The homepage content construction module configures content according to different areas of the homepage and combines multiple dimensions: user tags match personalized content according to user interests, behaviors and attributes; display duration distinguishes between long-term permanent content and short-term temporary content; scheduled display sets content to be displayed during specific time periods; the access platform adjusts the display content and format according to the device used by the user; when the user opens the application, the system dynamically generates and displays specific browsing content on the homepage based on pre-set rules and user tags.

[0064] In order to better illustrate the implementation method and system of the present invention for accurately delivering homepage recommended content based on user portrait, the following is a detailed description through specific embodiments. It should be noted that the following embodiments are only illustrative descriptions of the present invention, and the protection scope of the present invention is not limited thereto.

[0065] Example:

[0066] like Figure 1As shown, the present invention provides a method for accurately delivering homepage recommended content based on user portraits, including the following steps:

[0067] S1. The system comprehensively records various user behavior data during the user's use of the application, including but not limited to page browsing history, search history, click history, purchase history, collection history, and social interaction;

[0068] S2. Clean and denoise the collected raw data, filter out invalid and erroneous data, convert the format, standardize the data, ensure the accuracy of subsequent analysis, and store it in the database.

[0069] S3, the backend service generates multiple behavior tags for each user by analyzing the user's behavior data (e.g., browsing history, search history, click behavior, etc.) using predefined algorithms and rules. These tags describe the user's interests, preferences, and behavior patterns.

[0070] S4. The backend configures specific display content for each user tag and rule. These contents may include news articles, product information, advertisements, videos, social media content, etc. When a user comes in, the homepage specific browsing content is displayed according to the divided rules and tags.

[0071] In this embodiment, when a user browses a page, clicks a button, accesses a page, or performs a search operation, these behavior data will be sent to the server through an API request. After receiving the data, the server will filter it, remove invalid and erroneous data, and then conduct in-depth analysis and processing. Through predefined algorithms and rules, the user is analyzed in detail and multiple behavior tags are generated for each user. Based on these behavior tags, the system pre-configures specific homepage display content to achieve the effect of precise delivery.

[0072] In another embodiment, step S1 comprises:

[0073] Data collection includes defining the names of events that need to be collected, such as page browsing, search, click, purchase, collection, and social interaction, and defining the information content that needs to be collected for each event, such as page URL, search keyword, click button ID, purchase product ID, collection time, and social interaction type. When users trigger these events, the client records the relevant information and extracts the required data, encrypts the data using the RSA algorithm, and stores the encrypted data in the Redis cache for subsequent upload.

[0074] In another embodiment, step S2 comprises:

[0075] The client uploads the data in the cache to the server API interface in batches at fixed intervals. If the upload fails, the data will enter the retry queue to ensure that the upload is successful in the end. When the user closes the application, the client will re-upload the unsuccessful data at the next startup. After the server receives the uploaded data, it first verifies the data to ensure that the data has not been tampered with. Then, the data that passes the verification is decrypted and cleaned to filter out invalid and erroneous data. Next, the cleaned data is formatted and standardized, and the valid data is stored in the MongoDB database. Due to the large amount of user behavior data, the system uses Redis queues for consumption, obtains data from the queue, filters out invalid and erroneous data, and stores events in the MongoDB database. At the same time, data indexes are established in the database to improve the efficiency of data retrieval and analysis, and support complex query operations.

[0076] In another embodiment, step S3 comprises:

[0077] The backend service further cleans and denoises the data to ensure that each record is unique, and removes noise data such as invalid clicks and duplicate records to ensure data accuracy. The cleaned data is archived to ensure data integrity and traceability. The system uses predefined algorithms and rules to conduct in-depth analysis of user behavior data and generate multiple behavior tags for each user. These tags describe the user's interests, preferences, and behavior patterns. For example, if a user frequently browses the electronic product page, the system will generate an "interest tag: electronic products"; if the user searches for a specific brand multiple times, the system will generate a "brand preference tag." After the user behavior tag is generated, it is stored in the user portrait database, and the association between the tag and the user is established to ensure that the tag information of each user is complete. The tags are versioned and the generation time and update history of the tags are recorded.

[0078] In another embodiment, step S4 comprises:

[0079] Based on the user's behavior tags, the backend configures the corresponding display content, which can include news articles, product information, advertisements, videos, social media content, etc. For example, based on the user's "electronics" interest tag, the latest electronic product review articles are recommended. When a user accesses the app, the system dynamically adjusts the recommended content based on the user's behavior tags. By filtering and sorting, it ensures that the displayed content matches the user's current interests and needs, and displays the recommended content in a user-friendly way, such as displaying a list of recommended products and articles on the homepage.

[0080] The above-mentioned embodiments only express the specific implementation of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for implementing accurate delivery of recommended content on the homepage based on user portraits, characterized in that: The following steps are involved: S1. The system comprehensively records various user behavior data during the user's use of the application, including but not limited to page browsing history, search history, click history, purchase history, collection history and social interaction; S2. Clean and denoise the collected raw behavioral data, filter out invalid and erroneous data, convert the format, standardize the data, ensure the accuracy of subsequent analysis, and store it in the database; S3, the backend service generates multiple behavior tags for each user by analyzing the user's behavior data and using predefined algorithms and rules; these tags describe the user's interests, preferences, and behavior patterns; S4. The backend system configures content according to different regions of the homepage and multiple dimensions: user tags match personalized content based on user interests, behaviors, and attributes; display duration distinguishes between long-term resident content and short-term temporary content; and timed display sets content to be displayed during a specific time period; The access platform adjusts the display content and format according to the device used by the user; the display content includes news articles, product information, advertisements, videos, and social media content; When users use the application, specific browsing content on the homepage is displayed according to the divided rules and tags.

2. According to claim 1, a method for implementing accurate delivery of recommended content on the homepage based on user portraits is characterized in that: Step S1 includes: data collection includes defining the names of events that need to be collected, including page browsing, search, click, purchase, collection and social interaction, and defining the information content that needs to be collected for each event, including page URL, search keyword, click button ID, purchase product ID, collection time and social interaction type; when the user triggers these events, the client records the relevant information and extracts the required data, encrypts the data using the RSA algorithm, and stores the encrypted data in the Redis cache for subsequent upload.

3. According to claim 1, a method for implementing accurate delivery of recommended content on the homepage based on user portraits is characterized in that: Step S2 includes: User behavior data is consumed by Redis queues regularly to filter out invalid and erroneous data; data of different event types are classified and processed and stored in MongoDB for subsequent analysis and processing.

4. According to claim 1, a method for implementing accurate delivery of recommended content on the homepage based on user portraits is characterized in that: Step S3 includes: The backend service analyzes user behavior data and uses predefined algorithms and rules to generate multiple behavior tags for each user. These tags describe the user's interests, preferences, and behavior patterns. It uses machine learning algorithms to continuously optimize and update behavior tags to ensure the accuracy and real-time nature of recommended content.

5. According to claim 4, a method for implementing accurate delivery of homepage recommended content based on user portraits is characterized in that: Based on the recorded user behavior data, the system uses rule-based algorithms, statistical algorithms, or behavioral feature algorithms to generate labels.

6. A method for implementing accurate delivery of homepage recommended content based on user portraits according to claim 1, characterized in that: Step S4 includes: The backend configures specific display content for each user tag and rule, including news articles, product information, advertisements, videos, and social media content. When a user comes in, specific browsing content on the homepage is displayed according to the divided rules and tags. The system supports A / B testing and multiple recommendation strategies to evaluate and optimize recommendation effects and enhance user experience.

7. A system for accurately delivering recommended content on the homepage based on user portraits, characterized in that: Include: User behavior data collection module, which comprehensively records various user behavior data, including but not limited to page browsing history, search history, click history, purchase history, collection history and social interaction; The data preprocessing module is used to clean and denoise the collected raw data. The redis queue is used for consumption. The system obtains data from the queue, filters out invalid and erroneous data, and stores each event in the mongodb database for subsequent in-depth analysis and processing. Data indexes are established in mongodb to improve the efficiency of data retrieval and analysis, and support complex query operations. User tag building module, which analyzes user behavior data to build user tags. Based on the recorded user behavior data, the system uses predefined algorithms and rules to generate tags; The homepage content construction module configures content according to different regions of the homepage and multiple dimensions: user tags match personalized content based on user interests, behaviors, and attributes; display duration distinguishes between long-term resident content and short-term temporary content; and timed display sets content to be displayed during a specific time period; The access platform adjusts the display content and format according to the device used by the user; when the user opens the application, the system will dynamically generate and display specific browsing content on the homepage based on pre-set rules and user tags.

Images