Personalized service recommendation system based on user behavior analysis

By adopting a personalized recommendation system based on user behavior analysis in sports venues, combining geographical location, psychological state and consumption habits, providing personalized service recommendations with high accuracy, solving the lack of targeted and time-consuming service recommendations in the existing technology. The problem of this achieves in-depth exploration and immediate response to user needs, and improves user experience.

CN118964741AInactive Publication Date: 2024-11-15BEIJING DONGWANG TIANXIA TECH CO LTD

Patent Information

Application Number
CN202411148086.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The prior art is difficult to deeply understand users' immediate needs and potential preferences in sports venues, resulting in a lack of targeted and timely service recommendations.

Method used

A personalized recommendation system based on user behavior analysis is adopted, and personalized service recommendations are provided through data collection, real-time analysis and multimodal recommendation algorithms, combining geographical location, psychological state and consumption habits. The system includes a data collection unit, a data processing and analysis unit, a personalized recommendation unit, a cross-platform experience synchronization unit, and a security and privacy protection unit to ensure data security and user privacy.

Benefits of technology

It has achieved in-depth exploration and immediate response to the needs of sports venue users, provided high-accuracy personalized service recommendations, and improved user experience and satisfaction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118964741A_ABST
    Figure CN118964741A_ABST
Patent Text Reader

Abstract

The invention discloses a personalized service recommendation system based on user behavior analysis, and aims to improve user experience in environments such as stadiums. The system collects user behavior data by comprehensively using various sensors and intelligent equipment through the data collection unit, and performs deep analysis by using a machine learning algorithm through the data processing and analysis unit. And the personalized recommendation unit generates real-time recommendation to meet the instant demand of the user. The system also has a cross-platform experience synchronization function, and ensures that the user obtains consistent experience on different devices. In addition, the system pays attention to security and privacy protection, and the security of user data is ensured by adopting an advanced encryption technology and a privacy policy. Through intelligent prediction and optimization design of the user interface, highly personalized and user-friendly service recommendation is realized, and the method has a wide application prospect.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of customer relationship management (CRM), and in particular to a personalized service recommendation system based on user behavior analysis. Background Art

[0002] With the development of smart devices and big data technology, personalized services have gradually become a key factor in improving user experience. In environments such as sports stadiums, users' personalized needs for watching games and exercising are growing. Existing technologies usually use basic user preference settings or simple behavioral pattern recognition to provide service recommendations, but these methods often fail to deeply understand users' immediate needs and potential preferences, resulting in a lack of pertinence and timeliness in service recommendations. Summary of the invention

[0003] In order to solve the above problems, the present invention provides a personalized recommendation system based on behavioral analysis, which integrates multiple technologies such as cross-platform customer experience synchronization, personalized services based on geographic location, personalized services based on user psychological state, and intelligent prediction based on user consumption habits, so as to achieve in-depth mining and instant response to the needs of sports venue users.

[0004] To achieve the above objectives, this application adopts the following technical solutions:

[0005] Personalized service recommendation system, including:

[0006] A data collection unit, used to collect specific activity data of users in the sports stadium, with a collection frequency of at least once per second;

[0007] Data processing and analysis unit, used to analyze user behavior and psychological state, with analysis delay no more than 200 milliseconds after real-time data collection;

[0008] Personalized recommendation unit, used to generate personalized service recommendations based on user behavior, with the accuracy of the recommendation algorithm being at least 90%;

[0009] Cross-platform experience synchronization unit, used to synchronize user data to achieve a seamless cross-platform experience, with data synchronization time of no more than 1 second;

[0010] The security and privacy protection unit is used to ensure data security and user privacy, and the encryption algorithm used meets at least the AES-128 standard.

[0011] The data collection unit further includes: a camera component, which tracks user activity trajectories and records motion patterns, and recognizes at least 10 different motion patterns; an indoor positioning component, which realizes indoor positioning and user identity recognition, with a positioning accuracy of at least ±0.5 meters; a sensor monitoring component, which monitors the intensity and frequency of user activities, and recognizes at least 5 different activity intensity levels; a smart device data synchronization component, which synchronizes the user's physiological data and application usage, with a synchronization frequency of at least once a minute.

[0012] The data processing and analysis unit further includes: a data preprocessing component, which performs data cleaning, standardization and normalization, including at least missing value processing and outlier removal; a user behavior modeling component, including a K-means clustering algorithm unit, and the determination of the number of clusters k is based on the elbow rule and silhouette coefficient analysis; a psychological state assessment component, using a facial expression recognition unit, and identifying at least 5 basic emotional states; a consumer habit learning component, using a collaborative filtering algorithm unit, and the algorithm's recommendation similarity threshold is set to 0.7.

[0013] The personalized recommendation unit further includes: a multimodal recommendation algorithm unit, which generates recommendations based on user behavior, geographic location, psychological state and consumption habits, and the recommendation algorithm considers at least 3 different data sources; a real-time recommendation engine unit, which adopts online learning and incremental learning techniques, and the learning rate adjustment mechanism is based on dynamic evaluation of user feedback; a recommendation result explainability component, which provides reasons and basis for the recommendation, including at least the basic principles of the recommendation algorithm and the data types used.

[0014] The cross-platform experience synchronization unit further includes: a user account management component that implements secure identity recognition and data association through the OAuth 2.0 protocol and supports at least two different authentication methods; a device intelligent recognition and switching component that uses near-field communication technology to achieve seamless switching between devices with a switching time of no more than 0.5 seconds; and a cloud service data synchronization component that ensures centralized storage and real-time synchronization of data with a synchronization frequency of at least once every 5 seconds.

[0015] The geo-location-based service recommendations further include: an indoor positioning technology component that implements at least three different indoor positioning technologies, including Wi-Fi positioning, Bluetooth beacons, and UWB; a geographic information system (GIS) component that supports the storage and query of at least 10,000 geo-location data points; and an indoor navigation component that provides at least three different path planning algorithms, including the shortest path, the fastest path, and the path with the least congestion.

[0016] Service recommendations based on the user's psychological state further include: a physiological signal monitoring component that evaluates the user's emotional state such as heart rate, skin conductance, and breathing rate, and identifies at least three different physiological signal patterns; an emotional state assessment component, including facial expression recognition and voice emotion analysis technology, that identifies at least five different emotional states; an emotional intelligence feedback component that collects user feedback to optimize the recommendation algorithm, with feedback processing time not exceeding 10 seconds after real-time feedback.

[0017] Intelligent prediction based on user consumption habits further includes: a consumption habit identification component, which divides users into at least five different consumption habit groups through classification and cluster analysis; a prediction model component, which uses machine learning technologies such as random forests, gradient boosting machines (GBM), neural networks, etc., and includes at least 10 different prediction models; a user feedback loop component, which uses user acceptance, satisfaction and conversion rate feedback to optimize the recommendation model, and the feedback loop period does not exceed one week.

[0018] User interface and interaction design further includes: interface design components, which follow the principles of intuitiveness, consistency, and aesthetics, and support at least five different user interface themes; multimodal interaction components, including touch, voice, gesture, and visual recognition, and support at least ten different interaction methods; real-time feedback and user behavior tracking components, which are used to optimize recommendation algorithms and interface layouts, and the feedback processing time does not exceed 5 seconds after the real-time feedback.

[0019] The security and privacy protection unit further includes: data encryption components, such as SSL / TLS, to protect data transmission and storage security, and support at least two different encryption protocols; security authentication components, including multi-factor authentication and biometric technology, and support at least three different authentication technologies; privacy policy compliance components, which provide user data control rights, including access, correction and deletion, and the response time for data control operations shall not exceed 24 hours after the user's request.

[0020] Through the above technical solution, the personalized service recommendation system based on user behavior analysis of the present invention realizes in-depth mining and immediate response to the needs of sports venue users. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to make the embodiments, technical solutions and technical advantages of the present application clearer, the following drawings provide non-limiting views of the present application:

[0022] Figure 1 The overall architecture diagram of the personalized recommendation system provided by this application is shown.

[0023] Figure 2 A user behavior analysis process diagram is shown.

[0024] Figure 3A schematic diagram of the cross-platform customer experience synchronization process is shown.

[0025] Figure 4 A schematic diagram of the service recommendation process based on geographic location is shown.

[0026] Figure 5 A schematic diagram of the service recommendation process based on the user's psychological state is shown.

[0027] Figure 6 A schematic diagram of the intelligent prediction process based on user consumption habits is shown. DETAILED DESCRIPTION

[0028] In order to elaborate on the embodiments, technical solutions and technical advantages of the present application, a detailed and complete description will be provided below in conjunction with the accompanying drawings. It should be clear that the described embodiments only represent some application scenarios of the present application, not all possibilities. According to the embodiments of the present disclosure, those skilled in the art can obviously obtain other possible embodiments without performing creative work, and these embodiments also belong to the scope of protection of the present application.

[0029] 1. System deployment and data collection

[0030] The personalized recommendation system of the present invention is designed to provide a customized experience for each visitor of the stadium. The system is deployed in various key locations of the stadium, including the entrance area, seating area, catering service area, rest area, exercise facility area, etc. The deployed devices include but are not limited to cameras, Wi-Fi access points, Bluetooth beacons, pressure-sensitive carpets, motion trackers, etc., which together form a comprehensive data collection network.

[0031] Deploy high-definition cameras to capture the behavior patterns of users in the stadium. The cameras run at a frame rate of 30fps and have a resolution of 1920x1080 pixels. Use the YOLOv4 algorithm for object detection, and capture user behavior in real time through the camera unit. The confidence threshold is set to 0.6 and the NMS threshold is set to 0.4 to ensure the accuracy and real-time nature of behavior detection. For each detected object, the confidence of the bounding box is calculated using the formula:

[0032] Confidence=pobj×Π c∈classes p c

[0033] Indoor positioning is achieved by combining Wi-Fi access points and Bluetooth beacons. RSSI measurement technology is used to determine the user's location through trilateration. Kalman filter is used for state estimation, where the Q matrix represents process noise and the R matrix represents measurement noise, both of which are initialized to the unit matrix.

[0034] The system's data collection strategy emphasizes diversity and comprehensiveness. First, cameras and motion trackers collect users' activity trajectories and movement patterns in the venue, including walking speed, dwell time, activity area, etc. Second, Wi-Fi access points and Bluetooth beacons use signal strength changes to assist indoor positioning, while collecting the MAC address of the user's device to identify the user. In addition, pressure-sensitive carpets and motion trackers can monitor the intensity and frequency of users' activities in specific areas.

[0035] Smart devices carried by users, such as smartphones, smart watches, and health trackers, exchange data with the system through wireless networks. These devices are equipped with dedicated applications that can collect users' physiological data (such as heart rate, blood pressure, steps, etc.), application usage, and personal preferences. The system communicates with these devices through API interfaces to achieve automatic data synchronization.

[0036] The system design emphasizes the real-time and dynamic nature of data. Through real-time data stream processing technology, the system can instantly receive and process information from various data sources. This ensures the timeliness of recommended content and enables the system to quickly respond to changes in user needs. For example, when the system detects that a user has stayed in a certain area for too long, it can immediately recommend nearby rest areas or catering services.

[0037] During the data collection process, the system strictly adheres to the privacy protection principle. All collected data is anonymized to ensure that the user's identity is not identifiable. The system provides a transparent privacy policy, clearly informs users of the purpose, scope and use of data collection, and obtains the user's explicit consent. Users can view their own data at any time and modify or delete it.

[0038] The system uses a variety of security measures to protect the security of data. Encryption protocols such as SSL / TLS are used to protect data during transmission to prevent data from being intercepted and tampered with. In terms of data storage, the system uses encrypted storage technology to ensure that data stored on the server is not accessed by unauthorized persons. In addition, the system conducts security audits and vulnerability scans regularly to discover and fix potential security issues.

[0039] 2. Data Processing and Analysis

[0040] Data preprocessing is the first step in data analysis, which aims to ensure the quality and consistency of data. First, data cleaning is performed to identify and handle missing values, outliers, and duplicate records. Next, data transformation is performed, including data standardization and normalization, to eliminate the effects of different dimensions and magnitudes. In addition, data integration combines data from different sources to form a unified data view.

[0041] The collected physiological and behavioral data were preprocessed, including denoising and data standardization. A Gaussian filter was used with a standard deviation σ set to 3 to remove high-frequency noise in the signal. The calculation formula of the Gaussian filter kernel is:

[0042]

[0043] User behavior modeling uses machine learning techniques to analyze and predict user behavior in sports venues. The system uses unsupervised learning algorithms, such as K-means clustering, to classify user behavior and identify different user groups and behavior patterns. At the same time, supervised learning algorithms, such as support vector machines (SVM) and random forests, predict future behavior trends based on users' historical behavior.

[0044] The K-means clustering algorithm is used to classify user behaviors and identify different user behavior patterns. The algorithm initializes the number of centroids k to 5 and determines the optimal number of clusters through the elbow rule.

[0045] Geographic location service optimization provides users with the best service recommendations by analyzing users' real-time location data and combining it with the layout information of the venue. The system uses graph theory algorithms to calculate the shortest path and guide users to reach their destination quickly. In addition, spatial analysis technology is used to identify user-dense areas and provide venue managers with decision support for crowd control and resource allocation.

[0046] The psychological state assessment algorithm evaluates the user's emotional state by analyzing the user's physiological data and behavioral performance. The system uses emotional computing technology, combined with facial expression recognition, voice emotion analysis and physiological signal analysis, to comprehensively evaluate the user's emotional state. For example, by analyzing the user's facial expression and voice tone when watching a game, the system can determine whether the user is excited, nervous or relaxed.

[0047] The consumption habit learning mechanism is to learn the user's purchase preferences and consumption habits by analyzing the user's historical consumption records and online behaviors. The system uses collaborative filtering algorithms to find other users with similar behaviors to the target user and recommend products or services that these users like. At the same time, the content-based recommendation algorithm recommends similar products or services based on the user's previous consumption content.

[0048] The real-time recommendation engine is the core component of the system, responsible for converting analysis results into real-time recommendations. The engine uses online learning and incremental learning techniques to continuously update user models and recommendation strategies. When user behavior or preferences change, the recommendation engine can respond quickly and provide the latest recommendations. In addition, the engine uses the Multi-Armed Bandit algorithm to find a balance between exploring new recommendations and leveraging existing recommendations.

[0049] The data analysis module is designed to be highly scalable and flexible, able to adapt to stadiums of different sizes and different types of events. The module adopts a microservice architecture, and each analysis task can be deployed and expanded independently. In addition, the system supports plug-in algorithm integration, which can easily add new data analysis algorithms and models to meet changing business needs.

[0050] When conducting data analysis, the system strictly abides by data protection regulations and ethical standards. All data analysis activities undergo a data protection impact assessment (DPIA) to ensure that the impact on user privacy is minimized. The system also provides transparent data analysis reports to let users understand the process and results of data analysis.

[0051] 3. Personalized Recommendation Engine

[0052] The personalized recommendation engine adopts a layered architecture, including the data layer, model layer, logic layer and application layer. The data layer is responsible for storing user behavior data, item information, user portraits, etc. The model layer contains a variety of recommendation algorithm models, such as collaborative filtering, content recommendation, deep learning models, etc. The logic layer is responsible for processing the recommendation logic, such as fusing the output of different recommendation models and processing user feedback. The application layer is responsible for applying the recommendation results to the user interface and providing them to users.

[0053] In order to provide more accurate personalized recommendations, the system uses a multimodal recommendation algorithm that combines the user's behavioral data, geographic location, psychological state, and consumption habits. For example, the system not only considers the user's past purchases, but also the user's behavioral patterns and real-time location in the stadium, as well as the user's emotional state, and generates a recommendation list based on this information.

[0054] Combine the user's behavior data, geographic location, psychological state and consumption habits to generate personalized recommendations. Use the weighted sum formula S = ω 1 ·B+ω 2 ·G+ω 3 P, weight ω 1 ,ω 2 ,ω 3 Dynamically adjust based on user feedback.

[0055] User portrait is the core of the recommendation system. The system builds user portrait by analyzing user behavior data, preference settings, social network information, etc. User portrait includes not only basic information of users, such as age, gender, occupation, etc., but also user interests, behavior habits, emotional tendencies, etc. In the process of building user portrait, the system adopts incremental learning and online learning technology to continuously update user portrait to adapt to changes in user behavior.

[0056] When generating a recommendation list, the recommendation engine not only considers the user's historical preferences, but also introduces diversity and novelty. The system provides richer recommendation content by adjusting the parameters of the recommendation algorithm, such as lowering the user similarity threshold in collaborative filtering and increasing the proportion of novel items in the recommendation list. In addition, the system also adopts an exploratory recommendation strategy to regularly recommend items to users that are not completely consistent with their historical behavior in order to discover the user's potential interests.

[0057] In order to improve user trust in the recommendation system, the system focuses on the explainability of the recommendation results. When generating a recommendation list, the recommendation engine generates a recommendation explanation to explain the reasons and basis for the recommendation. For example, when recommending a product, the explanation can be "Based on your purchase history and the preferences of similar users, we recommend this product to you."

[0058] The system has designed a user feedback mechanism that allows users to evaluate the recommendation results, such as likes, disinterest, etc. User feedback is collected in real time and fed back to the recommendation engine to adjust the recommendation strategy. For example, if a user expresses disinterest in a recommendation, the system will reduce the frequency of similar recommendations.

[0059] In order to continuously optimize the recommendation effect, the system regularly conducts A / B testing to compare the effects of different recommendation algorithms or parameter settings. By analyzing user feedback on different recommendation lists, the system selects the optimal recommendation strategy. In addition, the system uses a multi-armed bandit algorithm to dynamically adjust the recommendation strategy and balance the relationship between exploration and utilization.

[0060] The recommendation engine is designed to be highly scalable and flexible, supporting rapid adjustment of recommendation strategies according to different business needs and scenarios. The system adopts a modular design, and different recommendation algorithms and components can be updated and replaced independently. In addition, the system supports custom recommendation algorithms, allowing developers to develop new recommendation models according to specific needs.

[0061] 4. Cross-platform experience synchronization

[0062] In order to achieve cross-platform experience synchronization, the system adopts a unified data management strategy to ensure the consistency of user data on all platforms. Whether the user interacts with the system through a smartphone, tablet, personal computer or an interactive screen in the venue, the user's behavior data, preferences and history are kept synchronized.

[0063] The system achieves identity recognition and data association through a user account system. Users create an account when they first use the system and register using their mobile phone number, email address or social media account. The system uses security authentication protocols such as OAuth 2.0 to ensure the security of the user login process. After the user logs in, the system uses the user's unique identifier, such as user ID or device ID, to associate the user's data across different devices and platforms.

[0064] The user account system is used to achieve identity recognition and data association, ensuring data consistency across platforms. OAuth2.0 protocol and JWT are used for secure identity authentication, and the token validity period is set to 1 hour.

[0065] The system is designed with intelligent recognition and switching mechanisms to enable users to switch seamlessly between different devices. When a user approaches or leaves a device, the system automatically recognizes the user's device through near-field communication technologies such as Wi-Fi, Bluetooth or NFC, and synchronizes the user's current status and preferences.

[0066] The system backend adopts cloud service architecture to achieve centralized storage and real-time synchronization of data. Each user's interaction and behavior data is uploaded to the cloud in real time, and high availability and consistency of data are ensured through distributed database and data synchronization technology. Cloud services also provide elastic computing resources and dynamically adjust resource allocation according to system load.

[0067] The front-end interface of the system adopts responsive design, which can adapt to the screen size and resolution of different devices. Using front-end technologies such as HTML5, CSS3 and JavaScript, the system interface can intelligently adjust the layout and elements to provide the best user experience. In addition, the system also provides personalized interface customization based on user habits and preferences.

[0068] The system provides a consistent user interface style and experience on different devices and platforms. The design team follows a unified design language and specifications, including color schemes, font styles, button controls, etc. This helps reduce the user's learning cost and improves user acceptance and satisfaction with the system.

[0069] The system not only maintains consistency visually, but also works hard on the consistency of user experience. Every user operation and feedback is recorded and learned by the system, and the system dynamically adjusts the recommended content and interface layout based on the user's usage habits and preferences. For example, if the user often checks the game schedule on a mobile device, the system will highlight the game schedule information on all devices.

[0070] The system is designed with a cross-platform function expansion mechanism to support the rapid expansion of new functions based on new device and platform features. The system backend provides an open API interface, allowing third-party developers and services to access and enrich the system's functions and services. At the same time, the system frontend supports plug-in expansion, which can quickly integrate new user interface components and interaction methods.

[0071] In the process of achieving cross-platform experience synchronization, the system always attaches importance to security and privacy protection. All data transmission uses encryption protocols such as TLS1.2 or higher. The system also provides data access control and user authorization mechanisms to ensure the security and privacy of user data. Users can view and manage their data usage at any time.

[0072] 5. Personalized services based on geographic location

[0073] The system uses advanced indoor positioning technologies to accurately capture the real-time location of users in the stadium. These technologies include but are not limited to Wi-Fi positioning, Bluetooth beacons, ultra-wideband (UWB), visual recognition, etc. Through these technologies, the system can achieve meter-level or even centimeter-level positioning accuracy in indoor environments.

[0074] The system integrates a geographic information system (GIS) to manage and analyze data related to geographic locations. GIS is used to store information such as map data of sports venues, facility layouts, user locations, and supports complex spatial queries and analysis, such as finding the nearest dining spots or calculating the shortest path from the user's current location to the exit.

[0075] Based on the user's real-time location information, the system provides location-aware service recommendations. For example, when a user is close to a sales point, the system can recommend the special snacks or drinks at that point. In addition, the system can also predict the service points that the user may be interested in based on the user's movement trajectory in the venue, and push relevant information in advance.

[0076] The system provides indoor navigation to help users find their destination quickly. Based on the user's current location and target location, the system can automatically plan the optimal route and provide information such as walking direction, distance, and estimated time of arrival. During navigation, the system can also dynamically adjust the recommended route based on real-time traffic and congestion.

[0077] When collecting and using location data, the system strictly adheres to the privacy protection principle. The user's location data is encrypted during transmission and storage to ensure data security. In addition, the system provides privacy setting options to allow users to control whether their location data is collected and how it is used.

[0078] The system supports personalized location-based service customization based on user preferences and historical behaviors. For example, if the system recognizes that a user is a vegetarian, when the user is close to a catering service point, the system will give priority to recommending vegetarian options. In addition, the system can also predict new service points that the user may be interested in based on the user's historical visit records and actively push relevant information.

[0079] When providing location-based services, the system integrates multi-dimensional information to enhance the accuracy and relevance of recommendations. This information includes user behavior patterns, preference settings, real-time crowd conditions, event arrangements within the venue, etc. By comprehensively considering these factors, the system can provide more accurate and personalized location services.

[0080] The system has the ability to monitor and analyze crowd flow in real time to optimize location services and improve user experience. By analyzing crowd density and flow direction, the system can identify crowded areas and adjust service recommendations in a timely manner to guide users to avoid peak areas. At the same time, crowd flow analysis results can also provide decision-making support for venue managers, such as adjusting resource allocation at service points or optimizing event arrangements.

[0081] The system is designed with an intelligent event trigger mechanism that automatically triggers related services based on the user's location changes. For example, when a user enters a stadium, the system automatically pushes a welcome message and today's event schedule; when a user approaches a specific area, the system pushes a detailed introduction to the area and recommended services. This intelligent trigger mechanism improves the initiative and timeliness of services.

[0082] 6. Personalized services based on user psychological state

[0083] The system uses a variety of psychological state monitoring technologies, including physiological signal monitoring, facial expression recognition and natural language processing. By monitoring the user's heart rate, skin conductance, breathing rate and other physiological signals, the system can assess the user's emotional state. Facial expression recognition technology is used to analyze the user's emotional changes, while natural language processing technology is used to analyze the user's language and tone.

[0084] The system has a built-in sentiment analysis model that uses machine learning algorithms to classify and analyze users' emotional tendencies. The model can identify users' emotional states, such as happiness, sadness, anger, or excitement, and adjust the recommendation strategy based on changes in emotional states.

[0085] The system combines the user's psychological state with behavioral data for in-depth analysis. By analyzing the user's behavioral patterns under different emotional states, the system can better understand the user's needs and preferences, thereby providing more intimate and personalized services.

[0086] Based on the user's emotional state, the system provides corresponding emotional response service recommendations. For example, if the system detects that the user is in a low mood, it can recommend relaxing music or video content, or provide some stress-relief and relaxation activities.

[0087] The system has the ability to dynamically track the psychological state, update the user's emotional state in real time, and adjust the recommended content in real time according to the emotional changes. This dynamic tracking ensures the timeliness and relevance of the recommendation service.

[0088] In the process of monitoring and analyzing the user's psychological state, the system strictly adheres to the principle of privacy protection. All data related to user emotions are encrypted, and the monitoring and analysis results of user emotions are only used to provide personalized services and will not be used for any other purpose.

[0089] The system is designed with an emotional intelligence feedback mechanism that allows users to provide feedback on the emotional tendencies of recommended content. User feedback is used to further train and optimize the sentiment analysis model, improving the accuracy of the model and the relevance of recommendations.

[0090] The system uses multimodal data fusion technology to identify the user's psychological state. Combining multiple data sources such as the user's physiological signals, facial expressions, voice intonation, etc., the system can provide a more accurate and comprehensive emotional state assessment.

[0091] The system supports personalized service customization based on the user's psychological state. Users can set the parameters of the emotional response service according to their needs and preferences, such as what kind of recommended content they want to receive in a specific emotional state.

[0092] In addition to providing immediate emotional response services, the system can also provide long-term emotional health management suggestions based on the user's emotional change trends. For example, if the system finds that the user is under long-term stress, it can recommend the user to participate in some relaxing and stress-relieving activities or courses.

[0093] 7. Intelligent prediction based on user consumption habits

[0094] The system collects and analyzes various types of consumption data of users, including but not limited to purchase history, browsing behavior, rating feedback and social interaction. By deeply mining these data, the system can identify the explicit and implicit consumption preferences of users.

[0095] Utilizing the user's behavioral sequences in sports venues, such as purchasing, renting or booking services, the system employs time series analysis and sequence pattern mining techniques to predict users' future consumption trends and demands.

[0096] The system has developed a special consumption habit recognition algorithm, which combines users' demographic information and behavioral data, and divides users into different consumption habit groups through classification and cluster analysis.

[0097] Machine learning techniques such as random forest, gradient boosting machine (GBM), neural network, etc. are used to build prediction models. The system continuously optimizes model parameters and improves prediction accuracy through user feedback and actual consumption behavior.

[0098] The LSTM network is used to model the user's consumption habits and predict future consumption trends. The LSTM network contains 3 hidden layers, 128 neurons in each layer, and the learning rate is set to 0.001.

[0099] Based on the predicted results of consumption habits, the system automatically generates personalized recommendations. These recommendations include not only the product or service itself, but also related discount information, usage suggestions and the best time to buy.

[0100] The system establishes a user feedback loop mechanism, and uses feedback information such as user acceptance, satisfaction and conversion rate to further train and adjust the prediction model to form a closed-loop optimization system.

[0101] The system evaluates the effectiveness of recommendation results from multiple dimensions, including coverage, novelty, surprise, and user satisfaction, to ensure the diversity and attractiveness of the recommended content.

[0102] The system intelligently selects the timing of recommendations based on the user's consumption habits and behavior patterns. For example, recommendations are provided during the time period when the user frequently consumes or when the user shows a purchase intention.

[0103] The system can generalize users’ consumption habits in different scenarios and realize personalized recommendations across scenarios. For example, users’ preferences on online shopping platforms can be used to enrich their recommendations in sports stadiums.

[0104] While providing consumption habit prediction services, the system also focuses on explainability and transparency. Users can clearly understand the logic behind the recommendations and increase their trust in the recommendation system.

[0105] When building and applying consumption habit prediction models, the system strictly adheres to data protection regulations and ethical standards to ensure that user privacy is not violated and that user choices and preferences are kept confidential.

[0106] 8. User Interface and Interaction Design

[0107] The user interface design follows the principles of intuitiveness, consistency and aesthetics. The system provides a clear visual hierarchy and intuitive user guidance to ensure that users can quickly understand how to interact with the system.

[0108] The system allows users to customize the interface according to their preferences and needs. Users can choose different themes, layouts and fonts, and even adjust the way recommended content is displayed.

[0109] User interface design follows the principles of intuitiveness, consistency, and aesthetics, and provides clear user guidance. The size and spacing of interface elements are designed according to Fitts' law to improve user interaction efficiency.

[0110] Recommended content is displayed through an interactive interface, and users can interact with the recommended content by clicking, sliding, or voice commands. The system responds to user interactions by providing more relevant information or adjusting recommended content.

[0111] Interface elements are designed to be context-sensitive, dynamically adjusting based on the user’s current behavior, location, and mental state. For example, when a user is watching a game, the interface may highlight game-related information and recommendations.

[0112] The system supports multimodal interaction, including touch, voice, gesture and visual recognition, which provides users with a rich interaction method to adapt to the habits and preferences of different users.

[0113] Taking into account the needs of different users, the system provides auxiliary functions such as voice reading, high contrast mode, and font enlargement to improve the accessibility of the system.

[0114] The user interface integrates a real-time feedback mechanism, allowing users to evaluate and provide feedback on recommended content. The system adjusts the recommendation strategy in real time based on user feedback to provide a more personalized experience.

[0115] The system integrates user behavior tracking capabilities at the user interface level to analyze user interaction patterns and preferences in order to optimize recommendation algorithms and interface layout.

[0116] In the user interface design, the system provides obvious privacy and security prompts to ensure that users understand how their data is collected and used, and can easily manage their privacy settings.

[0117] The system interface supports multiple languages ​​and can automatically switch languages ​​according to the user's geographic location and language preference to provide a localized user experience.

[0118] The user interface adopts responsive design and can adapt to different screen sizes and resolutions, providing a consistent user experience whether on mobile phones, tablets or desktop computers.

[0119] To help users better use the system, the interface provides user education and guidance functions, including tutorials, tips, and help documents, so that users can quickly master the functions of the system.

[0120] IX. System Security and Privacy Protection

[0121] The system uses industry-standard encryption technology to protect the security of user data. All sensitive data, including personal information and user behavior data, is encrypted during transmission and storage to ensure that the data will not be accessed by unauthorized persons at any time.

[0122] The AES-256 algorithm is used to encrypt user data to ensure the security of data during transmission and storage. The key length is 256 bits, using CBC mode, and the initialization vector (IV) length is 128 bits.

[0123] The system implements a multi-level security authentication mechanism, including but not limited to multi-factor authentication, biometric technology (such as fingerprint and facial recognition), and dynamic verification codes to ensure that only verified users can access their data and services.

[0124] The system deploys network security measures such as firewalls, intrusion detection systems and DDoS protection to prevent malicious attacks and unauthorized access. Network security policies are regularly updated to cope with the ever-changing network threats.

[0125] The system has formulated a clear privacy policy to clearly inform users how their data is collected, used and protected. Users must agree to the privacy policy before using the system, and the system provides convenient privacy settings, allowing users to understand and control the use of their data at any time.

[0126] Users have full control over their own data, including access, correction and deletion. The system provides a simple and easy-to-use interface that allows users to easily manage their data and privacy settings.

[0127] When designing and implementing new features or services, the system conducts a Privacy Impact Assessment (PIA) to ensure that the new features do not have adverse effects on user privacy.

[0128] When conducting data analysis and mining, the system uses data anonymization and de-identification technology to protect the identity of users from being disclosed. All data used for analysis is processed to ensure that it cannot be traced back to individuals.

[0129] The system complies with relevant data protection laws and regulations, such as the EU General Data Protection Regulation (GDPR). The system is designed with automated compliance checking tools to ensure that all data processing activities comply with legal requirements.

[0130] The system requires all employees and partners to receive privacy protection training and understand their responsibilities and obligations in protecting user privacy. Through regular training and assessment, the team's privacy awareness is improved.

[0131] The system has developed a detailed emergency response plan to deal with possible data leaks or other security incidents. The emergency response team is responsible for taking prompt action when an incident occurs to minimize the impact and protect the security of user data.

[0132] The system raises users' awareness of privacy protection through various channels, such as user interface prompts, educational articles and videos. The system encourages users to take active privacy protection measures, such as changing passwords regularly and using multi-factor authentication.

[0133] The system implements continuous security and privacy reviews, including internal audits and third-party security assessments. Through these reviews, the system is able to discover and fix potential security vulnerabilities, ensuring that the system's security and privacy protection are always in the best state.

Claims

1. A personalized service recommendation system based on user behavior analysis, characterized in that: include: - A data collection unit, used to collect specific activity data of users in the sports stadium, with a collection frequency of at least once per second; -Data processing and analysis unit, used to analyze user behavior and psychological state; -Personalized recommendation unit, used to generate personalized service recommendations based on user behavior; -Cross-platform experience synchronization unit, used to synchronize user data to achieve a seamless cross-platform experience; -Security and Privacy Protection Unit, used to ensure data security and user privacy.

2. The system according to claim 1, wherein the data collection unit further comprises: -The camera assembly tracks the user's activity trajectory and records the movement pattern; -The indoor positioning component realizes indoor positioning and user identification; -The sensing monitoring component monitors the intensity and frequency of user activities; -The smart device data synchronization component synchronizes the user's physiological data and application usage.

3. The system according to claim 1, wherein the data processing and analysis unit further comprises: -The data preprocessing component performs data cleaning, standardization and normalization, including at least missing value processing and outlier removal; - the user behavior modeling component includes a K-means clustering algorithm unit, and the number of clusters k is determined based on the elbow rule and silhouette coefficient analysis; - the mental state assessment component uses a facial expression recognition unit to identify basic emotional states; -The consumption habit learning component adopts a collaborative filtering algorithm unit.

4. The system according to claim 1, wherein the personalized recommendation unit further comprises: -The multimodal recommendation algorithm unit generates recommendations based on user behavior, geographic location, psychological state and consumption habits; -The real-time recommendation engine unit adopts online learning and incremental learning technology, and the learning rate adjustment mechanism is based on dynamic evaluation of user feedback; -The explainability component of the recommendation results provides reasons and basis for the recommendation, including at least the basic principles of the recommendation algorithm and the data types used.

5. The system according to claim 1, wherein the cross-platform experience synchronization unit further comprises: -The user account management component implements secure identity recognition and data association through the OAuth 2.0 protocol and supports at least two different authentication methods; -The device intelligent identification and switching component uses near field communication technology to achieve seamless switching between devices; -The cloud service data synchronization component ensures centralized storage and real-time synchronization of data.

6. The system according to claim 1, wherein the service recommendation based on geographic location further comprises: -The indoor positioning technology component implements at least three different indoor positioning technologies, including Wi-Fi positioning, Bluetooth beacon, and UWB; -The geographic information system (GIS) component supports storage and query of geographic location data points; -The indoor navigation component provides different path planning algorithms, including the shortest path, the fastest path and the path with the least congestion.

7. The system according to claim 1, wherein the service recommendation based on the user's psychological state further comprises: -The physiological signal monitoring component evaluates the user's emotional state such as heart rate, skin conductance, breathing rate, etc., and identifies different physiological signal patterns; - The emotional state assessment component includes facial expression recognition and voice emotion analysis technology to identify different emotional states; -The emotional intelligence feedback component collects user feedback to optimize the recommendation algorithm.

8. The system according to claim 1, wherein the intelligent prediction based on user consumption habits further comprises: -The consumption habit identification component divides users into at least five different consumption habit groups through classification and cluster analysis; -The prediction model component uses machine learning techniques such as random forest, gradient boosting machine (GBM), neural network, etc.; -The user feedback loop component uses user acceptance, satisfaction and conversion rate feedback to optimize the recommendation model, and the feedback loop period does not exceed one week.

9. The system according to claim 1, wherein the user interface and interaction design further comprises: -The interface design components follow the principles of intuitiveness, consistency, and aesthetics, and support different user interface themes; -The multimodal interaction components include touch, voice, gesture and visual recognition, supporting different interaction methods; -The real-time feedback and user behavior tracking components are used to optimize the recommendation algorithm and interface layout.

10. The system according to claim 1, wherein the security and privacy protection unit further comprises: -The data encryption component, such as SSL / TLS, protects data transmission and storage security and supports different encryption protocols; -The security authentication component includes multi-factor authentication and biometrics, supporting different authentication technologies; - The privacy policy follows the component, providing user data control rights, including access, correction and deletion.

Citation Information

Patent Citations

  • Time series database based cluster type body-building management system and implementation method thereof

    CN108538360A

  • System and device for detecting athletic performance of infant

    CN117695605A

  • Intelligent management system for sports field information

    CN118228846A

Cited By

  • AI psychological tutoring system with emotion intelligence and privacy calculation double engines

    CN120585332A