AI user portrait construction method and system based on biological characteristics

Abstract

The invention discloses an AI user portrait construction method and system based on biological characteristics, and the method comprises the steps: collecting and analyzing the physiological index data and behavior habit data of a user, recognizing the health state and behavior pattern of the user, constructing a personalized user portrait model, providing personalized suggestions for the user according to the user portrait model, and carrying out the recognition of the health state and behavior pattern of the user. Continuously collecting interactive feedback information of the user to the product or service, and optimizing the user portrait model; the construction system comprises a physiological index data acquisition module, a behavior habit data acquisition module, a data analysis processing module and a user interaction updating module. According to the method, the physiological index data and the behavior habit data of the user are collected and analyzed in real time, and the personalized user portrait model is constructed, so that the health requirement and the life style of the user can be known more deeply, more personalized product recommendation, content pushing, advertisement putting and other services are provided, and the participation degree, the satisfaction degree and the loyalty degree of the user can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and system for constructing an AI user portrait based on biometrics. Background Art

[0002] User portraits are digital descriptions of users that are constructed by analyzing their behaviors, preferences, and characteristics, thereby helping companies better understand user needs and optimize their products and services. With the rapid development of information technology and the widespread application of artificial intelligence (AI), user portrait technology has received increasing attention in all walks of life. However, traditional AI user portrait construction methods mainly rely on users' active behavior data, such as geographic location, browsing history, purchase history, and social media interactions.

[0003] The existing patent with publication number CN116484109A discloses an artificial intelligence-based customer portrait analysis system and method, which records the user's operation data on the web page through a crawling module, and describes the user's behavior through the construction and modification of primary and secondary attribute models. Although such portrait technology can reflect the user's spontaneous expectations and self-cognition, it may deviate from their actual needs due to the limitations of user cognition. For example, a user with high blood sugar should limit his carbohydrate intake, but the AI ​​user portrait method based on his previous shopping records may mark him as a sweet lover, and then recommend more high-sugar foods.

[0004] In recent years, the advancement of physiological feature recognition technology has made it possible to construct user portraits through the user's physiological indicators (such as blood sugar, blood pressure and heart rate). For example, the patent with publication number CN109767836A discloses a medical diagnosis artificial intelligence system, device and self-learning method thereof. The system collects and classifies raw data, uses data indicator algorithms to convert data, and then forms intuitive data, and performs medical model matching and risk reminders based on user portraits; however, this type of portrait technology based on user physiological data often focuses too much on application needs in the medical and health field, and still has certain limitations in combining with user behavior habits and other information. In addition, the portrait process mainly relies on physiological data passively provided by users, and rarely considers the user's subjective wishes and behavioral habits, resulting in users being unable to obtain comprehensive suggestions based on their health status and behavior patterns when using these systems, thereby affecting user experience and satisfaction. Summary of the invention

[0005] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0006] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a method and system for building an AI user portrait based on biometrics.

[0007] In order to achieve the above objectives, in a first aspect, the present invention provides a method for constructing an AI user portrait based on biometrics, comprising:

[0008] S100, collecting the user's physiological index data and behavioral habit data;

[0009] S200, using machine learning algorithms and data mining processes to analyze physiological indicator data and behavioral habit data;

[0010] Identify the user's health status and behavior patterns, build a personalized user portrait model, and provide personalized suggestions to users based on the user portrait model;

[0011] S300, continuously collects user feedback on product or service interactions and optimizes the user portrait model.

[0012] In some embodiments, the physiological index data include vital signs data, metabolic index data, body composition data and physiological rhythm data;

[0013] The behavior habit data includes eating behavior data, exercise behavior data and social behavior data;

[0014] The user's health status includes chronic disease risk level and potential health problem warnings, and the behavior pattern includes dietary preferences, exercise habits and social activities.

[0015] In some embodiments, step S100 includes:

[0016] S110, continuously reading and updating device drivers and adaptation protocols, adapting to different wearable devices or smart sensors, and collecting physiological indicator data of users;

[0017] S120, continuously acquiring multiple physiological indicator data, packaging, encrypting and synchronously transmitting the acquired physiological indicator data;

[0018] S130, establishing a multi-platform data connection interface to collect user behavior data based on behavior keywords, topic tags and behavior pattern rules;

[0019] S140, perceive and identify scene elements, and store and manage the collected behavior habit data based on timestamps.

[0020] In some embodiments, step S200 includes:

[0021] S210, identifying and eliminating erroneous or duplicated physiological indicator data and behavioral habit data, performing standardized integration, matching and association, and constructing a user data set;

[0022] S220, using machine learning algorithms and data mining processes to select user portrait feature subsets, and construct feature variables based on user portrait classification requirements;

[0023] S230, identifying the user's health status and behavior patterns, building a personalized user portrait and generating personalized recommendations.

[0024] In some embodiments, step S300 includes:

[0025] S310, divert and push personalized suggestions, and continuously collect user feedback on products or services;

[0026] S320, identify user consultation intentions, respond to communications, provide feedback on user satisfaction, and update user portrait labels.

[0027] In a second aspect, the present invention further provides an AI user portrait construction system based on biometrics, which is used to run the AI ​​user portrait construction method based on biometrics as described in the first aspect, and the construction system includes:

[0028] A physiological index data acquisition module is used to obtain the user's physiological index data through wearable devices or smart sensors;

[0029] Behavior habit data collection module, used to collect users' behavior habit data on different platforms;

[0030] The data analysis and processing module is used to analyze the collected physiological indicator data and physiological indicator data using machine learning algorithms and data mining processes, identify the user's health status and behavior patterns, and build a personalized user portrait;

[0031] The user interaction update module is used to provide users with personalized health management suggestions through personalized user portraits, continuously collect user interaction feedback information on products or services, and optimize the user portrait model.

[0032] In some embodiments, the physiological indicator data acquisition module includes:

[0033] The device adaptation sub-module is used to continuously read and update device drivers and adaptation protocols to adapt to different wearable devices or intelligent sensors;

[0034] The data acquisition and transmission sub-module is used to continuously obtain multiple physiological index data, package, encrypt and synchronously transmit the collected physiological index data.

[0035] In some embodiments, the behavior habit data acquisition module includes:

[0036] The multi-platform docking sub-module is used to establish a multi-platform data connection interface and collect user behavior data based on behavior keywords, topic tags and behavior pattern rules;

[0037] The scene perception sub-module is used to sense and identify scene elements and store and manage the collected behavior habit data based on timestamps.

[0038] In some embodiments, the data analysis and processing module includes:

[0039] The data cleaning and integration sub-module is used to identify and eliminate incorrect or duplicate physiological index data and behavior habit data, perform standardized integration, matching and association, and construct a user data set;

[0040] The feature training sub-module is used to screen the user portrait feature subset, construct feature variables based on user portrait classification requirements, use machine learning algorithms and data mining processes to identify the user's health status and behavior patterns, construct a personalized user portrait and generate personalized recommendations.

[0041] In some embodiments, the user interaction and update module includes:

[0042] The information push and collection sub-module is used to divert and push personalized recommendations and continuously collect the interactive feedback information of users on products or services;

[0043] The incentive and guidance sub-module is used to identify the user's consultation intention, respond to communication, feedback the user's satisfaction and update the user portrait tags.

[0044] The present invention has the following beneficial effects:

[0045] The present invention can effectively improve user engagement, satisfaction and loyalty by collecting and analyzing the user's physiological index data and behavior habit data in real time, combining the advantages of active and passive portrait technologies, constructing a personalized user portrait model, and providing more personalized product recommendations, content pushes, advertisement placements and other services to better understand the user's health needs and lifestyle. Description of the Drawings

[0046] Figure 1The process of the AI ​​user portrait construction method based on biometrics proposed by the present invention Figure 1 ;

[0047] Figure 2 The process of the AI ​​user portrait construction method based on biometrics proposed by the present invention Figure 2 ;

[0048] Figure 3 The process of the AI ​​user portrait construction method based on biometrics proposed by the present invention Figure 3 ;

[0049] Figure 4 The process of the AI ​​user portrait construction method based on biometrics proposed by the present invention Figure 4 ;

[0050] Figure 5 The principle block diagram of the AI ​​user portrait construction system based on biometrics proposed in the present invention;

[0051] Figure 6 The operation block diagram of the AI ​​user portrait construction system based on biometrics proposed in the present invention;

[0052] Figure 7 The process of the AI ​​user portrait construction method based on biometrics proposed by the present invention Figure 5 ;

[0053] Figure 8 The process of the AI ​​user portrait construction method based on biometrics proposed by the present invention Figure 6 ;

[0054] Fig. 9 The process of the AI ​​user portrait construction method based on biometrics proposed by the present invention Figure 7 ;

[0055] Fig.10 The process of the AI ​​user portrait construction method based on biometrics proposed by the present invention Figure 8 .

[0056] Legend:

[0057] 1. User portrait construction system; 11. Physiological indicator data collection module; 111. Equipment adaptation submodule; 112. Data collection and transmission submodule; 12. Behavioral habit data collection module; 121. Multi-platform docking submodule; 122. Scene perception submodule; 13. Data analysis and processing module; 131. Data cleaning and integration submodule; 132. Feature training submodule; 14. User interaction update module; 141. Information push collection submodule; 142. Incentive guidance submodule. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] The embodiment of the present application provides a method and system for constructing an AI user portrait based on biometrics, which solves the problem that the portrait technology based on user physiological data in the prior art often focuses too much on the application needs in the medical and health field, and still has certain limitations in combining with the user's behavioral habits and other information. In addition, the portrait process mainly relies on the physiological data passively provided by the user, and rarely considers the user's subjective wishes and behavioral habits, resulting in the user being unable to obtain comprehensive suggestions based on their health status and behavioral patterns when using these systems, thereby affecting the user experience and satisfaction. However, this application collects and analyzes the user's physiological indicator data and behavioral habit data in real time, combines the advantages of active and passive portrait technology, and constructs a personalized user portrait model, which can more deeply understand the user's health needs and lifestyle, and provide more personalized product recommendations, content push and advertising services, which can effectively improve user engagement, satisfaction and loyalty.

[0060] Please refer to the following examples for details:

[0061] Reference Figure 1-Figure 4 , the present invention provides an embodiment of a method for constructing an AI user portrait based on biometric features, the specific structure of which includes:

[0062] S100, collecting the user's physiological index data and behavioral habit data;

[0063] S200, using machine learning algorithms and data mining processes to analyze physiological indicator data and behavioral habit data;

[0064] Identify the user's health status and behavior patterns, build a personalized user portrait model, and provide personalized suggestions to users based on the user portrait model;

[0065] S300, continuously collects user feedback on product or service interactions and optimizes the user portrait model.

[0066] It should be explained in detail that the physiological index data include:

[0067] (1) Vital signs data, including heart rate, blood pressure, body temperature, etc., used to monitor cardiovascular health, exercise intensity adaptation, and emotional stress response;

[0068] (2) Metabolic indicator data, including blood sugar and blood lipids, can reflect the body's sugar metabolism level and thus adjust diet, exercise and drug treatment to prevent complications;

[0069] (3) Body composition data, including body fat percentage and muscle mass, which can reflect the body's basal metabolic rate, exercise capacity, physical recovery ability, etc.;

[0070] (4) Physiological rhythm data, including sleep rhythm and hormone secretion rhythm, which can monitor the time to fall asleep, sleep depth (light sleep, deep sleep, rapid eye movement duration), number of awakenings at night, etc., and evaluate changes in hormone rhythm.

[0071] Furthermore, behavioral habit data includes:

[0072] (1) Dietary behavior data, including food preference, dietary pattern, and food intake: This data can reflect dietary habits and potential risk of nutritional imbalance. Combined with the food nutrient database, it can estimate calorie and nutrient intake and provide a basis for personalized nutritional recommendations.

[0073] (2) Sports behavior data, including: exercise type preference, exercise frequency and exercise intensity, can be used to customize exercise plans, improve training effects and prevent sports injuries.

[0074] (3) Social behavior data, including: participation in social activities, social platform interactions, and communication styles. It can analyze content themes (such as healthy lifestyle sharing, food exploration, and exercise check-ins) to explore interests, hobbies, and values. Social platform data reflects users’ psychological needs and self-display tendencies, providing a basis for advertising push.

[0075] Correspondingly, the user's health status includes:

[0076] (1) Chronic disease risk level: including cardiovascular disease, diabetes, etc., determined by analyzing physiological data such as heart rate variability, blood pressure fluctuation patterns, and blood sugar metabolism indicators. For example, if a user's blood pressure is at a high level for a long time and their heart rate variability is low, then their risk level of cardiovascular disease may be high.

[0077] (2) Warning of potential health problems: including abnormal immune system or chronic fatigue syndrome, etc., which are predicted based on sleep quality data, self-assessed fatigue data, and inflammatory index data. For example, if the user's sleep quality is poor for a long time and he feels very tired, and the inflammatory index shows an upward trend, it may indicate the potential risk of abnormal immune system or chronic fatigue syndrome.

[0078] Furthermore, the behavior patterns include:

[0079] (1) Dietary preferences: including diet type, etc., which are divided into vegetarian diet, meat diet, high-protein diet, low-carb diet, etc. by analyzing the user's diet data over a period of time. If it is found that the user mainly consumes plant-based foods such as vegetables, fruits, and beans, and rarely eats meat, it can be judged that his diet type tends to be vegetarian.

[0080] (2) Exercise habits: including exercise cycle patterns and exercise type preferences, etc., are classified by analyzing the number of fixed exercises per week, exercise intervals, and exercise types. Assuming that a user regularly exercises on Monday, Wednesday, and Friday nights, then his exercise cycle pattern is three times a week, with one exercise per day. If a user frequently runs, swims, and other exercises, it can be judged that he prefers aerobic exercise.

[0081] (3) Social activities: including the frequency and types of participation in social activities. If a user attends multiple social gatherings per week and the types of gatherings are diverse, it means that his social activities are relatively active and the types of participation are diverse.

[0082] Please continue reading Figure 2 In this embodiment, step S100 includes:

[0083] S110, continuously reading and updating device drivers and adaptation protocols, adapting to different wearable devices or smart sensors, and collecting physiological indicator data of users;

[0084] S120, continuously acquiring multiple physiological indicator data, packaging, encrypting and synchronously transmitting the acquired physiological indicator data;

[0085] S130, establishing a multi-platform data connection interface to collect user behavior data based on behavior keywords, topic tags and behavior pattern rules;

[0086] S140, perceive and identify scene elements, and store and manage the collected behavior habit data based on timestamps.

[0087] For example, by regularly connecting to the official device manufacturer's website, technical forums, and device update push channels, the latest device drivers and adaptation protocols are continuously read. Once a new driver or protocol version is released, it is automatically downloaded and updated to the system; then, based on the updated device information library, the corresponding driver and adaptation protocol are accurately matched, and the connection parameters are automatically adjusted to ensure that a stable and efficient data transmission link is established with the device; then, multiple physiological indicator data are continuously acquired, and multi-channel synchronous acquisition technology is used to ensure that all physiological indicator data are collected at the same time point to avoid data synchronization problems caused by time differences; different types of physiological indicator data are then packaged and encrypted in a specific format to prevent the data from being stolen or tampered with during transmission, and transmitted to subsequent processing links in real time; based on pre-set behavioral keywords, topic tags, and behavioral pattern rules, the data is stored and analyzed in real time, and each piece of data is given a timestamp accurate to milliseconds to analyze the sequence of user behaviors, periodic laws, and the relationship between behaviors.

[0088] Please continue reading Figure 3 In this embodiment, step S200 includes:

[0089] S210, identifying and eliminating erroneous or duplicated physiological indicator data and behavioral habit data, performing standardized integration, matching and association, and constructing a user data set;

[0090] S220, using machine learning algorithms and data mining processes to select user portrait feature subsets, and construct feature variables based on user portrait classification requirements;

[0091] S230, identifying the user's health status and behavior patterns, building a personalized user portrait and generating personalized recommendations.

[0092] For example, by setting a reasonable physiological range threshold and combining the time series continuity of the data, the changing trend of the data is analyzed to see if it conforms to normal physiological laws to identify abnormal data; through the text duplication detection algorithm, the content posted by users on different platforms is compared to identify identical or highly similar duplicate records. For the identified errors and duplicate data, the data cleaning and integration submodule removes them from the original data set to ensure the accuracy and validity of the data; all data are normalized and mapped to the interval of [0,1], and the physiological indicator data and behavioral habit data are accurately matched and associated based on the user's unique identifier (such as user ID). By establishing a data association table, the physiological indicator data of the same user at different time points are integrated with the corresponding behavioral habit data to construct a comprehensive and systematic user data set.

[0093] The feature selection algorithm based on information gain is adopted. This algorithm evaluates the importance of each feature by calculating the information gain value of each feature to the target variable (such as the user's health risk category, consumption preference type, etc.); combined with statistical methods such as chi-square test, the correlation between each feature and the target variable is further evaluated. By setting a strict threshold, the features with significant correlation with the target variable are screened out, and those redundant and irrelevant features that contribute less to the user profile are removed, thereby reducing the data dimension and improving the efficiency and accuracy of model training; based on the different classification requirements of user profiles, such as health status classification, consumption behavior classification, interest and hobby classification, etc., the feature engineering submodule manually constructs a series of targeted feature variables; using the processed user data set and carefully selected feature subsets, a variety of machine learning and deep learning algorithms are used to accurately identify the user's health status and behavior patterns.

[0094] Based on the recognition results of the user's health status and behavior patterns, a personalized user portrait is constructed. The user portrait presents the various characteristics and attributes of the user in an intuitive way, including basic information (age, gender, occupation, etc.), health status (chronic disease risk level, current health status, etc.), behavior patterns (dietary preferences, exercise habits, social activity patterns, etc.) and interests and hobbies; based on the personalized user portrait, personalized recommendations are generated for the user. In terms of health management, targeted dietary adjustment suggestions, exercise plans and regular physical examination reminders are provided for users with higher health risks; in terms of consumption recommendations, relevant products and services are recommended based on the user's interests and behavior patterns.

[0095] Please continue reading Figure 4 In this embodiment, step S300 includes:

[0096] S310, divert and push personalized suggestions, and continuously collect user feedback on products or services;

[0097] S320, identify user consultation intentions, respond to communications, provide feedback on user satisfaction, and update user portrait labels.

[0098] For example, based on the diverse features presented by the user portrait, such as the user's health status, consumption preferences, interests and hobbies, and daily behavioral habits, the push strategy is intelligently planned to achieve efficient diversion of personalized suggestions; for health management suggestions, precise push is carried out according to the health risk level: if the user is judged to be at high risk of cardiovascular disease, the system will give priority to sending urgent and important health tips through the SMS channel, including detailed dietary adjustment suggestions, such as reducing the intake of high-salt and high-fat foods, and increasing the intake of vegetables and fruits rich in dietary fiber, and attaching professional exercise plans. These personalized suggestions will be pushed again in the form of pop-ups and message lists in the user's commonly used health management mobile applications. In terms of consumption recommendations, targeted push notifications are made based on the user's consumption behavior patterns and interests. For users who are keen on outdoor sports, the system will push suitable outdoor equipment through in-site messages during the shopping period on the e-commerce platform. At the same time, it will push product recommendations and activity information related to outdoor sports in related sports apps. The system collects user feedback information, uses natural language processing technology to conduct in-depth analysis of the feedback text, extracts key information, emotional tendencies and potential needs from user feedback through lexical analysis, syntactic analysis and semantic understanding, and adjusts personalized suggestions and product recommendations to ensure user participation and satisfaction in the health management process.

[0099] Reference Figure 5-Figure 6 The present invention also provides an embodiment of an AI user portrait construction system based on biometrics, which is used to run the AI ​​user portrait construction method based on biometrics in the above embodiment, and the construction system includes:

[0100] A physiological index data acquisition module is used to obtain the user's physiological index data through wearable devices or smart sensors;

[0101] Behavior habit data collection module, used to collect users' behavior habit data on different platforms;

[0102] The data analysis and processing module is used to analyze the collected physiological indicator data and physiological indicator data using machine learning algorithms and data mining processes, identify the user's health status and behavior patterns, and build a personalized user portrait;

[0103] The user interaction update module is used to provide users with personalized health management suggestions through personalized user portraits, continuously collect user interaction feedback information on products or services, and optimize the user portrait model.

[0104] Please continue reading Figure 5-Figure 6 In this embodiment, the physiological index data acquisition module includes:

[0105] (1) Device adaptation submodule, used to continuously read and update device drivers and adaptation protocols to adapt to different wearable devices or smart sensors;

[0106] By continuously updating device drivers and adaptation protocols, the system can ensure seamless connection with emerging smart wearable products, such as smart clothing and smart glasses. This submodule has a built-in device identification database. When a new device is connected, it can quickly identify its type, brand, and model, and automatically match the optimal connection parameters and data transmission format to ensure the smooth start of data collection; and personalized calibration plans can be formulated for different devices to unify data standards. For example, for heart rate monitoring bracelets of different brands, due to differences in sensor accuracy and algorithms, the calibration submodule uses authoritative medical equipment as a benchmark to perform real-time correction on the collected heart rate data to ensure that the data accuracy is at the same comparable level, providing a reliable basis for subsequent comprehensive analysis.

[0107] (2) a data acquisition and transmission submodule, which is used to continuously acquire multiple physiological index data, package, encrypt and synchronously transmit the collected physiological index data;

[0108] Vital signs such as heart rate, blood pressure, and blood sugar are measured through the integration of a variety of high-precision sensor chips. Each sensor is equipped with an independent signal conditioning circuit to amplify and filter weak physiological signals, remove environmental noise and interference, such as motion artifacts and electromagnetic interference, and accurately extract pure physiological indicator signals; and can use multi-channel synchronous acquisition technology to obtain multiple physiological indicator data at the same time, avoiding data asynchrony caused by acquisition time difference, and ensuring that various physiological data reflect the physical state at the same time. For example, in one heartbeat cycle, the electrocardiogram signal, heart sound signal, and fingertip pulse wave signal are synchronously collected to capture the heart function state in all directions, providing rich information for cardiovascular system health assessment.

[0109] Please continue reading Figure 5-Figure 6 In this embodiment, the behavior habit data collection module includes:

[0110] (1) A multi-platform docking submodule, which is used to establish a multi-platform data connection interface and collect user behavior data based on behavior keywords, topic tags, and behavior pattern rules;

[0111] By establishing a wide range of platform connection interfaces, it is able to achieve safe and compliant data docking with mainstream social platforms, e-commerce platforms, sports and health software, etc., and automatically update adaptation strategies in response to changes in API rules on different platforms to ensure stable capture of user behavior data on each platform; and it has a data screening function, which can accurately screen out valid information related to user eating, exercise, social, entertainment and other behavioral habits from massive platform data based on pre-set behavioral keywords, topic tags and behavioral pattern rules, providing data for subsequent in-depth analysis.

[0112] (2) Scene perception submodule, which is used to perceive and identify scene elements and store and manage the collected behavior habit data based on timestamps;

[0113] Using computer vision technology, we can intelligently analyze the pictures and videos uploaded by users to social platforms, cloud photo albums, etc. We can identify scene elements in the picture through deep learning models, such as sports equipment and venues in sports scenes, food types and tableware in eating scenes, judge the user's ongoing behavior activities, assist in collecting behavior habit data, and make up for the lack of information actively input by users; and we can give accurate timestamps to various types of behavior habit data collected, and store and manage them in an orderly manner according to the time series, so as to lay out personalized health management and product recommendations in advance and accurately capture the dynamic evolution characteristics of user behavior habits.

[0114] Please continue reading Figure 5-Figure 6 In this embodiment, the data analysis and processing module includes:

[0115] (1) Data cleaning and integration submodule, which is used to identify and eliminate erroneous or duplicated physiological indicator data and behavioral habit data, perform standardized integration, matching and association, and construct a user data set;

[0116] For the raw data collected from the physiological index data collection module and the behavioral habit data collection module, data cleaning algorithms, such as rule-based outlier detection and duplicate data removal algorithms, are used to identify and remove obviously erroneous or duplicate data records. For example, outrageous physiological values ​​caused by temporary sensor failures and the same user behavior information captured repeatedly during multi-platform collection processes are removed to ensure the accuracy and uniqueness of the data; then, data from different sources and formats are standardized and integrated, and the time format and coding method of the data are unified, so that physiological data and behavioral data can be accurately matched and associated according to individual users, and a complete user data set is constructed.

[0117] (2) Feature training submodule, which is used to screen user profile feature subsets, construct feature variables based on user profile classification requirements, use machine learning algorithms and data mining processes to identify users' health status and behavior patterns, build personalized user profiles, and generate personalized recommendations;

[0118] Use automatic feature selection techniques, such as statistical methods based on information gain and chi-square test, to screen out feature subsets that have key distinguishing and predictive power for building user portraits from user data sets; and through a variety of machine learning and deep learning algorithm frameworks, including random forests, support vector machines, linear regression, deep learning neural networks, etc., flexibly select and switch algorithms according to different portrait construction tasks (such as health status assessment, shopping preference analysis); establish an independent test data set, regularly conduct a comprehensive evaluation of the trained user portrait model, and quantify the accuracy, stability and generalization ability of the analysis model in predicting user health status, behavioral trends, etc. based on evaluation indicators. For example, when analyzing user health risks, it is found that certain specific dietary patterns and exercise frequency combinations can more effectively predict cardiovascular disease risks than other features. These features are selected first for model training to reduce data redundancy and improve model training efficiency. For large-scale user data, a distributed training architecture is used to divide the data into multiple subsets and assign them to different computing nodes for parallel training to accelerate model convergence. Combined with pre-set model evaluation indicators (such as accuracy, recall, F1 value, etc.), the optimal parameter combination is continuously sought during model training to ensure that the model performance is optimal and accurately portrays user characteristics and behavior patterns. Once a decline in model performance is found (such as a sudden drop in accuracy), the early warning mechanism is triggered in a timely manner to trace back and analyze the reasons, which may be changes in data distribution, model overfitting, etc., and then targeted optimization measures are taken, such as re-collecting and updating data, adjusting the model structure, to ensure that the portrait model continues to operate efficiently.

[0119] Please continue reading Figure 5-Figure 6 In this embodiment, the user interaction update module includes:

[0120] (1) Information push and collection submodule, which is used to push personalized suggestions and continuously collect user feedback on products or services;

[0121] Integrate multiple notification channels within mobile applications, intelligently divert and push information based on its urgency and importance, and accurately customize the theme, language style, and presentation of the pushed content based on the user portrait model. For example, for emergency information such as critical blood sugar values ​​and sudden health risk warnings, SMS messages are prioritized to reach users and emergency contacts in real time to ensure timely delivery of information; general product recommendations and health tips are pushed through in-app messages to avoid excessively disturbing users. Push illustrated sports product recommendation information with trendy elements to young, fashionable, and fitness-conscious users; for elderly users, use large fonts and concise language for health tips to fit the reading habits and preferences of different user groups and improve information acceptance.

[0122] (2) The incentive guidance submodule is used to identify the user's consultation intention, respond to communication, feedback user satisfaction, and update user portrait labels;

[0123] According to the user's health status and portrait characteristics, we set up phased health management goals for users, such as weight loss goals, blood sugar control range, etc., and track the progress in real time, and show it to users in a visual way, such as progress bars, charts, etc. When users achieve their goals, we give them timely rewards and encouragement to stimulate their inner motivation and encourage them to move towards a better health state, while providing dynamic data for continuous optimization of the system.

[0124] Please continue reading Figure 7-Figure 10 , which are specific embodiments of the AI ​​user profile construction method and system that combine physiological indicator data and behavioral habit data. These embodiments demonstrate different application scenarios of the system and how to build and update user profiles through real-time data collection and analysis.

[0125] Specific embodiment 1: construction and interaction of personalized user portraits of ordinary users:

[0126] In the modern consumer environment, using AI profiles of users to make personalized product recommendations can greatly improve user experience and satisfaction. By combining users' physiological data (such as blood sugar, blood pressure and heart rate) and living habits data to create AI profiles of users, the system can not only provide users with purchase recommendations based on their daily habits, but also adjust the recommendation strategy in a timely manner when abnormal health indicators are detected, and issue health tips to users.

[0127] Implementation steps include:

[0128] (1) User information collection:

[0129] Users register through a mobile application, enter basic information (such as age, gender, weight, height, etc.) and lifestyle habits (such as eating preferences, exercise habits); users wear wearable devices (such as smart watches, bracelets or non-invasive smart sensors), which monitor the user's physiological indicators in real time, especially blood sugar, blood pressure and heart rate.

[0130] (2) Physiological data monitoring:

[0131] The system collects the user's blood sugar, blood pressure and heart rate data in real time through wearable devices, and transmits the data to the data analysis and processing module; the system sets health thresholds, monitors changes in the user's physiological indicators in real time, and records the time and value of each measurement.

[0132] (3) User behavior data records:

[0133] The system records the user's eating habits, including: types of food for each meal (such as high-sugar, high-fat foods), food intake (such as grams or number of servings), eating times (such as the specific times of breakfast, lunch and dinner) and eating frequency (such as the number of meals and snacks per day); the system records the user's exercise habits, including: daily exercise types (such as aerobic exercise, strength training, etc.), duration of each exercise (such as minutes), total daily activity (such as steps or calories consumed); the system records the user's living habits, including: sleep time and quality (such as time to fall asleep, time to wake up, number of awakenings at night), stress level (such as user-rated stress level), social activities (such as social gatherings attended, frequency of interaction with friends).

[0134] (4) User portrait construction:

[0135] The data analysis and processing module combines the user's physiological indicator data (such as blood sugar, blood pressure, heart rate, etc.) with behavioral habit data (such as diet records, exercise volume), uses machine learning algorithms (such as random forests, support vector machines, linear regression and deep learning neural networks) to analyze the data and build a personalized user portrait model for the user; system analysis aspects include: health status (analyzing the stability of blood sugar levels, blood pressure changes and heart rate conditions, and using regression analysis to determine the user's health risks) and shopping preferences (identifying the types and brands of food that users frequently purchase, consumption frequency and amount, and using cluster analysis to summarize users' consumption behaviors).

[0136] (5) Personalized recommendation generation:

[0137] When the user's health indicators are normal, the system recommends related products based on the user's shopping preferences and consumption habits, such as frequently purchased health foods, sports equipment or leisure activity-related products; when it detects that the user's physiological indicators are abnormal (such as high blood sugar levels), the system adjusts the recommendation strategy and promotes health-related products, such as low-sugar foods, healthy recipe books or related health monitoring equipment; at the same time, the system also issues health tips, suggesting that users pay attention to their diet or increase exercise to improve their health.

[0138] (6) User Interaction:

[0139] The system pushes personalized purchase recommendations and health tips to users through mobile applications and provides real-time feedback options; users can record daily diet, exercise and physiological data (such as blood sugar, blood pressure and heart rate) in the application, and the system updates the user portrait in real time; user feedback (such as purchasing behavior, health changes) will be uploaded in real time, and the system will continuously optimize user portraits and recommendation strategies based on these feedbacks.

[0140] (7) Continuous monitoring and optimization:

[0141] The system regularly analyzes users' health and consumption data to identify their shopping trends and potential health risks. Based on the users' latest feedback and physiological data, the system adjusts personalized recommendations to ensure users' participation and satisfaction in the consumption and health management process.

[0142] Through this embodiment, the system can effectively build a personalized user portrait model for ordinary users, and improve the user's shopping experience and health management effect through real-time interaction; users can not only obtain targeted product recommendations, but also adjust their shopping and health management strategies in real time through the system's feedback mechanism, so as to better achieve consumption and health goals; the system monitors physiological indicators such as blood sugar, blood pressure and heart rate to help users understand the correlation between these indicators and personal living habits and consumption behaviors, so as to provide more targeted personalized services.

[0143] Specific Example 2: Construction and Interaction of AI User Profiles for Users Who Want to Lose Weight

[0144] With the improvement of health awareness, more and more people hope to achieve weight loss goals through scientific methods. Through AI user portrait technology, the system can combine the user's physiological data (such as blood sugar, blood pressure and heart rate) and behavioral habit data to provide users with personalized weight loss suggestions, product recommendations and interactive experiences.

[0145] Implementation steps include:

[0146] (1) User information collection:

[0147] Users register through a mobile application, enter basic information (such as age, gender, weight, height, etc.), weight loss goals (such as expected weight, weight loss time) and lifestyle habits (such as eating preferences, exercise habits); users wear wearable devices (such as smart watches, bracelets or non-invasive smart sensors), which monitor the user's physiological indicators in real time, especially blood sugar, blood pressure and heart rate.

[0148] (2) Physiological data monitoring:

[0149] The system collects the user's blood sugar, blood pressure and heart rate data in real time through wearable devices, and transmits the data to the data analysis and processing module; the system sets health thresholds, monitors changes in the user's physiological indicators in real time, and records the time and value of each measurement.

[0150] (3) User behavior data records:

[0151] The system records the user's eating habits, including: food types for each meal (such as high-sugar, high-fat foods), food intake (such as grams or number of servings), eating times (such as the specific times of breakfast, lunch and dinner) and eating frequency (such as the number of meals and snacks per day); the system records the user's exercise habits, including: daily exercise types (such as aerobic exercise, strength training, etc.), the duration of each exercise (such as minutes) and total daily activity (such as steps or calories consumed); the system records the user's living habits, including: sleep time and quality (such as time to fall asleep, time to wake up, number of awakenings at night) and stress level (such as user-rated stress level).

[0152] (4) AI user portrait construction:

[0153] The data analysis and processing module combines the user's physiological data (such as blood sugar, blood pressure, heart rate, etc.) with behavioral data (such as diet records, exercise volume), uses machine learning algorithms (such as random forests, support vector machines and deep learning neural networks) to analyze the data and build a personalized user portrait model; system analysis aspects include: health status (analyzing the impact of blood sugar levels, blood pressure changes and heart rate on weight loss, and using regression analysis and classification algorithms to determine the user's health risks), eating habits (identifying the relationship between the intake of high-sugar and high-fat foods and weight changes through cluster analysis) and exercise habits (analyzing the impact of daily activity and exercise type on heart rate and blood sugar, and using time series analysis methods to monitor exercise effects).

[0154] (5) Personalized suggestions and product recommendations:

[0155] Based on the AI ​​user portrait, the system analyzes the user's weight loss progress and generates personalized weight loss management suggestions and product recommendations, such as: Dietary suggestions: recommend low-sugar, low-fat recipes, and remind users to control calorie intake; Exercise suggestions: recommend suitable aerobic exercise and strength training, and provide a weekly exercise plan to help users increase their basal metabolic rate; Product recommendations: promote products related to weight loss, such as health foods, fitness equipment and nutritional supplements.

[0156] (6) Adjustment strategy: When weight loss progress is not ideal: If the user's weight loss progress does not meet expectations, the system will re-evaluate the user's diet and exercise plan, increase exercise intensity or adjust diet structure, and provide incentives such as health challenges; When user indicators are abnormal: If the user's blood sugar, blood pressure or heart rate levels are abnormal, the system will automatically trigger an alarm and recommend that the user immediately adjust the diet (such as reducing the intake of high-sugar foods) or increase rest time. At the same time, the user may be advised to consult a medical professional.

[0157] (7) User Interaction:

[0158] The system pushes personalized weight loss management advice and product recommendations to users through mobile applications, and provides real-time feedback options; users can record daily diet, exercise and physiological data (such as blood sugar, blood pressure and heart rate) in the application, and the system updates the AI ​​user portrait in real time. User feedback (such as weight changes, diet records) will be uploaded in real time, and the system will continuously optimize the AI ​​user portrait and recommendation strategy based on these feedbacks.

[0159] (8) Continuous monitoring and optimization:

[0160] The system regularly analyzes users' health data to identify their weight loss trends and potential risks. Based on the users' latest feedback and physiological data, the system adjusts personalized suggestions and product recommendations to ensure users' participation and satisfaction in the weight loss process.

[0161] Through this embodiment, the system can effectively build personalized AI user portraits for users who want to lose weight, and improve the user's weight loss effect through real-time interaction. Users can not only get targeted diet and exercise advice, but also adjust their weight loss strategies in real time through the system's feedback mechanism, so as to better achieve their weight loss goals and improve their quality of life. By monitoring physiological indicators such as blood sugar, blood pressure and heart rate, the system helps users understand the correlation between these indicators and weight loss, and provides more targeted personalized services and product recommendations.

[0162] Specific Example 3: AI user portrait construction and product recommendation for high blood sugar / diabetes users:

[0163] Diabetic patients need to carefully manage their blood sugar levels in daily life. Through AI user profiling technology, the system can combine the user's physiological data (such as blood sugar, heart rate, blood pressure) and behavioral habit data to provide personalized health advice and product recommendations to help users better manage their condition.

[0164] Implementation steps include:

[0165] (1) User information collection:

[0166] Users register through a mobile app and enter basic information (such as age, gender, weight, height, etc.) and health status (such as diabetes type, treatment plan). Users wear wearable devices (such as smart watches, bracelets or blood glucose monitors), which monitor the user's physiological indicators in real time, especially blood sugar, heart rate and blood pressure.

[0167] (2) Physiological data monitoring: The system collects the user's blood sugar, heart rate and blood pressure data in real time through wearable devices and transmits the data to the data analysis and processing module. The system sets health thresholds, monitors the changes in the user's physiological indicators in real time, and records the time and value of each measurement.

[0168] (3) User behavior data records:

[0169] The system records the user's eating habits, including: food types (such as carbohydrate content), food intake (such as grams or number of servings) for each meal, meal times (such as the specific times of breakfast, lunch and dinner) and meal frequency (such as the number of meals and snacks per day); the system records the user's exercise habits, including: daily exercise types (such as walking, swimming, yoga, etc.), duration of each exercise (such as minutes) and total daily activity (such as number of steps or calories consumed).

[0170] (4) AI user portrait construction:

[0171] The data analysis and processing module combines the user's physiological data (such as blood sugar, heart rate, etc.) with behavioral data (such as diet records, exercise volume), uses specific machine learning algorithms (such as random forests, support vector machines, linear regression and deep learning neural networks) to analyze the data and build a personalized user portrait model for the user; system analysis aspects include: health status (analyzing the frequency and amplitude of blood sugar fluctuations to identify potential health risks), eating habits (monitoring the intake of high-carbohydrate foods to assess their impact on blood sugar levels) and exercise habits (such as the impact of daily activity and exercise type on blood sugar).

[0172] (5) Personalized suggestions and product recommendations:

[0173] Based on the AI ​​user portrait, the system analyzes the user's health status and generates personalized health management suggestions and product recommendations, such as: Dietary suggestions: recommend low-sugar, high-fiber recipes, and remind users to properly control carbohydrate intake; Exercise suggestions: recommend suitable aerobic exercise and strength training, and provide a weekly exercise plan to help users stabilize blood sugar; Product recommendations: promote products related to diabetes management, such as blood glucose monitors, low-sugar foods, diabetes-specific nutritional supplements, and healthy recipe books.

[0174] (6) Adjustment strategy: Optimization of blood sugar management: If the user's blood sugar fluctuates greatly, the system will recommend adjustments to diet and exercise plans, and may recommend the use of more frequent blood sugar monitoring equipment; Abnormal user indicators: When the user's blood sugar or heart rate levels are abnormal, the system will automatically trigger an alarm, suggesting that the user immediately adjust their diet or rest, and may suggest that the user contact a medical professional for further guidance.

[0175] (7) User Interaction:

[0176] The system pushes personalized health management suggestions and product recommendations to users through mobile applications, and provides real-time feedback options; users can record daily diet, exercise and physiological data (such as blood sugar, heart rate) in the application, and the system updates the AI ​​user portrait in real time; user feedback (such as blood sugar changes, diet records) will be uploaded in real time, and the system will continuously optimize the AI ​​user portrait and recommendation strategy based on these feedback.

[0177] (8) Continuous monitoring and optimization: The system regularly analyzes users’ health data to identify their health trends and potential risks. Based on the users’ latest feedback and physiological data, the system adjusts personalized suggestions and product recommendations to ensure users’ participation and satisfaction in the health management process.

[0178] Through this embodiment, the system can effectively build personalized AI user portraits for users with high blood sugar or diabetes, and improve the user's health management effect through real-time interaction. Users can not only get targeted diet and exercise advice, but also adjust their health management strategies in real time through the system's feedback mechanism, so as to better control blood sugar levels and improve the quality of life. By monitoring physiological indicators such as blood sugar, heart rate and blood pressure, the system helps users understand the correlation between these indicators and health management, and provides more targeted personalized services and product recommendations.

[0179] Specific embodiment 4: AI user portrait construction and product promotion for elderly users:

[0180] Elderly users often face challenges from various chronic diseases, such as hypertension and coronary heart disease. Through AI user profiling technology, the system can combine the user's physiological data (such as blood pressure, heart rate) and living habits to provide personalized health advice and product recommendations, thereby improving user experience and health management effects.

[0181] Implementation steps include:

[0182] (1) User information collection:

[0183] Elderly users register through a mobile app and enter basic information (such as age, gender, height, weight, etc.) and health conditions (such as history of hypertension and coronary heart disease). Users wear wearable devices (such as smart watches, bracelets or medical-grade health monitoring devices), which monitor the user's physiological indicators in real time, especially blood pressure and heart rate;

[0184] (2) Physiological data monitoring: The system collects the user's blood pressure and heart rate data in real time through wearable devices and transmits the data to the data analysis and processing module; the system sets health thresholds, monitors the changes in the user's physiological indicators in real time, and records the time and value of each measurement.

[0185] (3) User behavior data records:

[0186] The system records the user's eating habits, including: food types for each meal (such as low-salt foods, foods rich in omega-3 fatty acids), food intake (such as grams or servings), meal times (such as the specific times of breakfast, lunch and dinner) and meal frequency (such as the number of meals and snacks per day); the system records the user's exercise habits, including: daily exercise types (such as walking, Tai Chi, yoga, etc.), duration of each exercise (such as minutes) and total daily activity (such as number of steps or calories consumed).

[0187] (4) AI user portrait construction:

[0188] The data analysis and processing module combines the user's physiological data (such as blood pressure, heart rate, etc.) with behavioral data (such as diet records, exercise volume), uses specific machine learning algorithms (such as random forests, support vector machines, linear regression and deep learning neural networks) to analyze the data and build a personalized user portrait model for the user; system analysis aspects include: health status (assessing the user's cardiovascular health risk by analyzing the frequency and amplitude of blood pressure fluctuations and the stability of heart rate, and using regression analysis to predict potential health problems), eating habits (monitoring the intake of low-salt diets and heart-healthy foods, using classification algorithms to identify the user's eating patterns, and providing users with targeted dietary advice) and exercise habits (analyzing the user's daily activity volume and the impact of different types of exercise (such as walking, Tai Chi, yoga) on cardiovascular health, and using cluster analysis methods to identify the most suitable exercise plan for the user).

[0189] (5) Personalized suggestions and product recommendations:

[0190] Based on the AI ​​user portrait, the system analyzes the user's health status and generates personalized health management suggestions and product recommendations, such as: Dietary suggestions: recommend low-salt, fiber- and antioxidant-rich recipes, and remind users to increase their vegetable and fruit intake; Exercise suggestions: recommend suitable low-intensity aerobic exercise and provide a weekly exercise plan to help users improve their cardiovascular health; Product recommendations: promote products related to the management of hypertension and coronary heart disease, such as blood pressure monitors, ECG monitoring devices, heart health supplements, and healthy recipe books.

[0191] (6) Adjustment strategy:

[0192] Health management optimization: If the user's blood pressure or heart rate fluctuates greatly, the system will recommend adjusting the diet and exercise plan, and may recommend the use of health monitoring equipment with higher frequency; Abnormal user indicators: When the user's blood pressure or heart rate levels are abnormal, the system will automatically trigger an alarm, suggesting that the user immediately adjust his or her lifestyle (such as reducing salt intake or increasing rest), and may suggest that the user contact a medical professional for further guidance.

[0193] (7) User Interaction:

[0194] The system pushes personalized health management suggestions and product recommendations to users through mobile applications, and provides real-time feedback options; users can record daily diet, exercise and physiological data (such as blood pressure, heart rate) in the application, and the system updates the AI ​​user portrait in real time; user feedback (such as health changes, product usage experience) will be uploaded in real time, and the system will continuously optimize the AI ​​user portrait and recommendation strategy based on this feedback.

[0195] (8) Continuous monitoring and optimization:

[0196] The system regularly analyzes users' health data to identify their health trends and potential risks. Based on the users' latest feedback and physiological data, the system adjusts personalized suggestions and product recommendations to ensure users' participation and satisfaction in the health management process.

[0197] Through this embodiment, the system can effectively build personalized AI user portraits for elderly users with common underlying diseases such as hypertension and coronary heart disease, and improve the user's health management effect and product promotion effect through real-time interaction. Users can not only get targeted diet and exercise advice, but also adjust their health management strategies in real time through the system's feedback mechanism, so as to better control blood pressure and heart health and improve the quality of life. By monitoring physiological indicators such as blood pressure and heart rate, the system helps users understand the correlation between these indicators and health management, and provides more targeted personalized services and product recommendations.

[0198] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing an AI user portrait based on biometrics, characterized in that: include: S100, collecting the user's physiological index data and behavioral habit data; S200, using machine learning algorithms and data mining processes to analyze physiological indicator data and behavioral habit data; Identify the user's health status and behavior patterns, build a personalized user portrait model, and provide personalized suggestions to users based on the user portrait model; S300, continuously collects user feedback on product or service interactions and optimizes the user portrait model.

2. The method for constructing an AI user portrait based on biometrics according to claim 1, characterized in that: The physiological index data include vital signs data, metabolic index data, body composition data and physiological rhythm data; The behavior habit data includes eating behavior data, exercise behavior data and social behavior data; The user's health status includes chronic disease risk level and potential health problem warnings, and the behavior pattern includes dietary preferences, exercise habits and social activities.

3. The method for constructing an AI user portrait based on biometrics according to claim 1, characterized in that: Step S100 includes: S110, continuously reading and updating device drivers and adaptation protocols, adapting to different wearable devices or smart sensors, and collecting physiological indicator data of users; S120, continuously acquiring multiple physiological indicator data, packaging, encrypting and synchronously transmitting the acquired physiological indicator data; S130, establishing a multi-platform data connection interface to collect user behavior data based on behavior keywords, topic tags and behavior pattern rules; S140, perceive and identify scene elements, and store and manage the collected behavior habit data based on timestamps.

4. The method for constructing an AI user portrait based on biometrics according to claim 1, characterized in that: Step S200 includes: S210, identifying and eliminating erroneous or duplicated physiological indicator data and behavioral habit data, performing standardized integration, matching and association, and constructing a user data set; S220, using machine learning algorithms and data mining processes to select user portrait feature subsets, and construct feature variables based on user portrait classification requirements; S230, identifying the user's health status and behavior patterns, building a personalized user portrait and generating personalized recommendations.

5. The method for constructing an AI user portrait based on biometrics according to claim 1, characterized in that: Step S300 includes: S310, divert and push personalized suggestions, and continuously collect user feedback on products or services; S320, identify user consultation intentions, respond to communications, provide feedback on user satisfaction, and update user portrait labels.

6. A biometric-based AI user portrait construction system, characterized in that: The construction system is used to run the AI ​​user portrait construction method based on biometrics according to any one of claims 1 to 5, and the construction system includes: A physiological index data acquisition module is used to obtain the user's physiological index data through wearable devices or smart sensors; Behavior habit data collection module, used to collect users' behavior habit data on different platforms; The data analysis and processing module is used to analyze the collected physiological indicator data and physiological indicator data using machine learning algorithms and data mining processes, identify the user's health status and behavior patterns, and build a personalized user portrait; The user interaction update module is used to provide users with personalized health management suggestions through personalized user portraits, continuously collect user interaction feedback information on products or services, and optimize the user portrait model.

7. The AI ​​user portrait construction system based on biometrics according to claim 6 is characterized in that: The physiological index data acquisition module includes: The device adaptation submodule is used to continuously read and update device drivers and adaptation protocols to adapt to different wearable devices or smart sensors; The data acquisition and transmission submodule is used to continuously acquire multiple physiological index data, package, encrypt and synchronously transmit the collected physiological index data.

8. The AI ​​user portrait construction system based on biometrics according to claim 6, characterized in that: The behavior habit data collection module includes: The multi-platform docking submodule is used to establish a multi-platform data connection interface and collect user behavior data based on behavior keywords, topic tags and behavior pattern rules; The scene perception submodule is used to perceive and identify scene elements and store and manage the collected behavior habit data based on timestamps.

9. The AI ​​user portrait construction system based on biometrics according to claim 6, characterized in that: The data analysis and processing module comprises: The data cleaning and integration submodule is used to identify and eliminate erroneous or duplicated physiological indicator data and behavioral habit data, perform standardized integration, matching and association, and construct a user data set; The feature training submodule is used to screen user portrait feature subsets, build feature variables based on user portrait classification requirements, use machine learning algorithms and data mining processes to identify users' health status and behavior patterns, build personalized user portraits, and generate personalized recommendations.

10. The AI ​​user portrait construction system based on biometrics according to claim 6, characterized in that: The user interaction update module includes: The information push collection submodule is used to push personalized suggestions and continuously collect user feedback on products or services. The incentive guidance submodule is used to identify user consultation intentions, respond to communications, provide feedback on user satisfaction, and update user portrait labels.

Images