Personalized exercise evaluation system

Abstract

The invention discloses a personalized exercise evaluation system, which is characterized in that the system performs accurate modeling on multi-dimensional features of a user on the basis of introduction of a user portrait technology, supports dynamic updating, realizes accurate user grouping by utilizing a classification model, and performs exercise evaluation according to a classification result and a user portrait. The system provides personalized exercise health assessment and suggestion reports for users and comprises a control center module, the control center module is connected with a data acquisition and preprocessing module, a user portrait generation module, a classification model operation and user grouping module and a personalized health assessment and recommendation module, and all the modules are in control connection with the control center module. According to the method and the system, user portrait construction is combined with a classification model, user multi-dimensional features are accurately modeled, the generalization ability of the system is improved, and the method and the system are suitable for personalized features and requirements of different user groups.

Description

Technical Field

[0001] The present invention relates to the technical field of health exercise evaluation systems, and in particular to a personalized exercise evaluation system. Background Art

[0002] With the development of artificial intelligence technology, various fields are vigorously using and developing applications. For the sports health assessment of personnel, most of them rely on self-input single-function devices without classification models. Some designs have classification models for sports health assessment, and some try to use machine learning technology for sports health assessment. This type of solution collects a certain amount of user data and trains a simple classification model to predict the user's health status. The main performance is as follows:

[0003] Data collection: Combine questionnaires and equipment to obtain users' basic health and exercise data.

[0004] Model training: Based on the labeled dataset, the model is trained to predict the user's exercise adaptability (such as high-intensity adaptation, medium-intensity adaptation).

[0005] Evaluation results: Recommend exercise plans to users based on model classification results.

[0006] Technical features:

[0007] A classification model was introduced, but the model was limited to simple scenarios, lacked comprehensive modeling of complex user characteristics, and had poor generalization ability.

[0008] Summary of technical details of existing technologies

[0009]

[0010] The shortcomings of these existing technical solutions are precisely the background for the present invention, which solves the above problems in a targeted manner by combining the technical means of user portrait and classification model, while balancing the system versatility and personalized needs. Summary of the invention

[0011] The purpose of the present invention is to provide a personalized sports assessment system, characterized in that the system is based on the introduction of user portrait technology, accurately models the multi-dimensional characteristics of users, supports dynamic updates, uses classification models to achieve accurate user grouping, and provides users with personalized sports health assessment and recommendation reports based on classification results and user portraits. The system includes a control center module, the control center module is connected to a data acquisition and preprocessing module, a user portrait generation module, a classification model operation and user grouping module, and a personalized health assessment and recommendation module, and each module is controlled and connected to the control center module respectively, and the data acquisition module, the user portrait generation module, the classification model operation and user grouping module, and the personalized health assessment and recommendation module are associated in sequence, and the associated information is transmitted to each other.

[0012] The data acquisition and preprocessing module is connected to the wearable device and the user terminal device, and collects user data in real time through the Bluetooth or Wi-Fi protocol, and completes the cleaning and normalization operations through the preprocessing module. The normalization formula is:

[0013]

[0014] Among them, x is the original data and x′ is the normalized data.

[0015] The user portrait generation module extracts the static and dynamic features of the user through feature engineering based on the results of the data collection and preprocessing module, uses feature vectors to represent the user portrait, and maps the user features to a high-dimensional space. The feature extraction is:

[0016] 1) Average heart rate:

[0017] 2) Exercise frequency: count the number of days per week for exercise;

[0018] 3) Feature vector representation:

[0019] v = [age, gender, BMI, exercise frequency, average heart rate, ...]

[0020] The classification model operation and user grouping module are connected to a support vector machine SVM, which uses a machine learning algorithm to classify user portraits to form a model, and is trained through historical annotation data to generate classification rules. The classification function algorithm of the support vector machine is:

[0021] f(x)=sign(w·x+b)

[0022] Among them, w is the weight vector, x is the user feature vector, and b is the bias.

[0023] The personalized health assessment and recommendation module is designed with a health assessment formula:

[0024] Health score = w1·BMI+w2·heart rate stability+w3·exercise frequency

[0025] Generate a personalized health assessment report based on user grouping results, historical data and evaluation formulas, and recommend an exercise plan suitable for the user, including exercise type, intensity and frequency.

[0026] Generate exercise plan matching: recommend appropriate exercise intensity based on the user's health score.

[0027] The data acquisition and preprocessing module design value input code is:

[0028]

[0029] The user portrait generation module is designed based on the data acquisition and preprocessing module and the input code is:

[0030] The classification model operation and user grouping module design input code is:

[0031] The personalized health assessment and recommendation module design input code is:

[0032]

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The present invention realizes a personalized exercise evaluation system, and based on the introduction of user portrait technology, accurately models the multi-dimensional characteristics of users, supports dynamic updating, and uses classification models to realize accurate user grouping. According to the classification results and user portraits, personalized health exercise evaluation and recommendation reports are provided to users, thereby realizing personalized health exercise evaluation for different user groups on the basis of universality, and at the same time realizing the combination of user portrait construction and classification model and accurate modeling of user multi-dimensional characteristics, thereby improving the generalization ability of the system, adapting to the characteristics and needs of different user groups, supporting multi-source data fusion, comprehensively analyzing the user's exercise data, physical characteristics and health status, dynamically updating the evaluation results, and adapting to changes in user status in real time, which is very suitable for large-scale user groups. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a schematic diagram of the system architecture of the present invention.

[0036] Figure 2This is a flow chart of the application of the system of the present invention.

[0037] Figure 3 A more detailed flow chart is applied to the system of the present invention. DETAILED DESCRIPTION

[0038] 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.

[0039] Definition or brief explanation of technical keywords

[0040] 1. User Profile

[0041] Definition: By analyzing the user's multi-dimensional data (including physical characteristics, exercise habits, health status, etc.), a comprehensive feature model is generated to describe the user's behavior, needs and health status.

[0042] Function: User portrait is the basis for personalized health assessment and supports the input and dynamic update of classification models.

[0043] 2. Classification Model

[0044] Definition: A model that uses machine learning technology to group users. By analyzing user portrait data, users are divided into different categories or groups (such as "high-intensity adaptation group" and "low-intensity adaptation group").

[0045] Function: The classification model supports accurate matching of personalized exercise recommendations and improves the scientificity and applicability of evaluation results.

[0046] 3. Dynamic Feedback

[0047] Definition: By collecting new user data (such as heart rate, number of steps, etc.) in real time, the user portrait and classification model are dynamically adjusted to ensure that the evaluation results can reflect the user's latest status.

[0048] Function: Improve the real-time and accuracy of health assessment and exercise recommendations.

[0049] 4. Multi-Source Data Fusion

[0050] Definition: Integrate and process multiple user data sources (such as wearable device data, questionnaire data, medical records, etc.) to generate comprehensive user information through feature extraction and modeling.

[0051] Role: Ensure the comprehensiveness and multi-dimensional analysis capabilities of health assessment.

[0052] 5. Health Assessment Report

[0053] Definition: A report generated based on user portraits and classification models, including the user's current health status, exercise adaptability analysis, and personalized exercise recommendations.

[0054] Function: Provide personalized health assessment results to enhance users’ awareness of their health status and their motivation to improve health behaviors.

[0055] 6. Wearable Devices

[0056] Definition: Smart devices used to collect users' real-time exercise and health data, such as smart bracelets, smart watches, etc.

[0057] Function: It is an important source of real-time data collection for the system and supports dynamic updating of user portraits.

[0058] 7. Cloud Platform

[0059] Definition: A distributed computing architecture responsible for data storage, processing, and model execution, providing support for system efficiency and scalability.

[0060] Function: To realize multi-user access and large-scale data processing, and support real-time calculation and distribution of evaluation results.

[0061] 8. Feature Engineering

[0062] Definition: In the process of building a user portrait, effective features that can reflect the user's health status and exercise needs are extracted and optimized from the original data.

[0063] Function: Improve the training efficiency and evaluation accuracy of classification models.

[0064] 9. Real-Time Monitoring

[0065] Definition: Through wearable devices or other sensors, the user's motion and physiological data, such as heart rate, number of steps, calorie consumption, etc., are collected in real time.

[0066] Function: Support dynamic feedback mechanism to ensure the timeliness of system evaluation results.

[0067] 10. Visualization

[0068] Definition: Use charts, text, and other forms to intuitively present health assessment reports and exercise recommendations to users.

[0069] Role: Improve the interpretability and user acceptance of evaluation results.

[0070] 11. Supervised Learning

[0071] Definition: A machine learning technology that trains classification models by annotating data to predict and group users’ health status.

[0072] Function: It is the core technology of the classification model and provides algorithm support for health assessment.

[0073] 12. Exercise Adaptability

[0074] Definition: Assess the appropriate intensity, type, and frequency of exercise based on the user's physical characteristics and health status.

[0075] Function: It is an important basis for personalized exercise recommendations.

[0076] Abbreviations

[0077] ML (Machine Learning) definition: Building classification models through data-driven methods for prediction and clustering.

[0078] SVM (Support Vector Machine) definition: A classification model that can effectively process the multi-dimensional features of user health data and achieve accurate clustering.

[0079] BMI (Body Mass Index) definition: A health indicator calculated by height and weight, used to assess the user's body shape and obesity level.

[0080] BPM (Beats Per Minute)

[0081] Definition: Indicates the user's heart rate, often used to assess exercise intensity and health status.

[0082] Example 1

[0083] See also Figure 1 to Figure 2The present invention provides a personalized sports assessment system, characterized in that the system is based on the introduction of user portrait technology, accurately models the multi-dimensional features of users, supports dynamic updates, uses classification models to achieve accurate user grouping, and provides users with personalized sports health assessment and recommendation reports based on classification results and user portraits. The system includes a control center module, the control center module is connected to a data acquisition and preprocessing module, a user portrait generation module, a classification model operation and user grouping module, and a personalized health assessment and recommendation module, and each module is controlled and connected to the control center module respectively, and the data acquisition module, the user portrait generation module, the classification model operation and user grouping module, and the personalized health assessment and recommendation module are sequentially associated and transmit associated information to each other.

[0084] The data acquisition and preprocessing module is connected to the wearable device and the user terminal device, and collects user data in real time through the Bluetooth or Wi-Fi protocol, and completes the cleaning and normalization operations through the preprocessing module. The normalization formula is:

[0085]

[0086] Among them, x is the original data and x′ is the normalized data.

[0087] The user portrait generation module extracts the static and dynamic features of the user through feature engineering based on the results of the data collection and preprocessing module, uses feature vectors to represent the user portrait, and maps the user features to a high-dimensional space. The feature extraction is:

[0088] 1) Average heart rate:

[0089] 2) Exercise frequency: count the number of days per week for exercise;

[0090] 3) Feature vector representation:

[0091] v = [age, gender, BMI, exercise frequency, average heart rate, ...]

[0092] The classification model operation and user grouping module are connected to a support vector machine SVM, which uses a machine learning algorithm to classify user portraits to form a model, and is trained through historical annotation data to generate classification rules. The classification function algorithm of the support vector machine is:

[0093] f(x)=sign(w·x+b)

[0094] Among them, w is the weight vector, x is the user feature vector, and b is the bias.

[0095] The personalized health assessment and recommendation module is designed with a health assessment formula:

[0096] Health score = w1·BMI+w2·heart rate stability+w3·exercise frequency

[0097] Generate a personalized health assessment report based on user grouping results, historical data and evaluation formulas, and recommend an exercise plan suitable for the user, including exercise type, intensity and frequency.

[0098] Generate exercise plan matching: recommend appropriate exercise intensity based on the user's health score.

[0099] The data acquisition and preprocessing module design value input code is:

[0100]

[0101] The user portrait generation module is designed based on the data acquisition and preprocessing module and the input code is:

[0102] The classification model operation and user grouping module design input code is:

[0103]

[0104] The personalized health assessment and recommendation module design input code is:

[0105] Example 2

[0106] The process of implementing the technical solution using the above system

[0107] Objective: Personalized sports health assessment based on user portrait and classification model

[0108] Step 1: Data collection and preprocessing

[0109] Executing entities: user's wearable devices (such as smart bracelets) and user terminal devices (such as smartphones). Triggering conditions: the user starts wearing the device and completes the pairing of the device and the phone, and starts sports data monitoring.

[0110] Action Description:

[0111] Wearable devices collect users’ real-time exercise data (such as heart rate, number of steps, calorie consumption, exercise duration, etc.).

[0112] The user terminal is connected to the wearable device via Bluetooth, and the data is received and cached locally.

[0113] The user terminal uploads the collected data to the cloud platform via the HTTPS protocol.

[0114] The cloud platform cleans and preprocesses the data (such as removing noise and filling missing values).

[0115] The result: cleaned multi-dimensional user data, including physical characteristics, exercise habits and real-time health status.

[0116] Function: Provide accurate data foundation for subsequent user portrait modeling and classification model.

[0117] Step 2: User portrait generation

[0118] Execution entity: User portrait modeling module in the cloud server.

[0119] Trigger condition: The user completes at least one complete data upload.

[0120] Action Description:

[0121] The system extracts features from the user's basic data (such as height, weight, age, etc.).

[0122] Combined with real-time monitoring data (such as heart rate fluctuations and exercise frequency), the user's exercise preference and health status feature vector is generated.

[0123] User portraits include static features (such as gender and age) and dynamic features (such as current heart rate and sleep status).

[0124] The result: a dynamically updated, multi-dimensional user profile that comprehensively reflects the user’s health status and exercise preferences.

[0125] Function: Provide input data for the classification model and improve the personalization ability of the system.

[0126] Step 3: Classification model operation and user grouping

[0127] Execution entity: Classification model in the cloud server.

[0128] Trigger condition: The system completes the generation or update of the user portrait.

[0129] Action Description:

[0130] The system groups users according to their portraits using a trained classification model (such as “high-intensity exercise adaptation group” and “low-intensity exercise adaptation group”).

[0131] The classification model uses the support vector machine (SVM) algorithm, inputs the user's feature vector, and outputs the classification result.

[0132] The system assigns the corresponding exercise plan template to the user based on the classification results.

[0133] The result: users are accurately classified into groups suitable for a certain intensity and type of exercise.

[0134] Function: Through classification results, the user groups can be refined to lay the foundation for subsequent personalized recommendations.

[0135] Step 4: Personalized health assessment and recommendations

[0136] Execution entity: Recommendation system module in the cloud server, combined with the display module of the user terminal. Trigger condition: The classification model completes the grouping of users.

[0137] Action Description:

[0138] The system combines the user classification results with historical data to generate a health assessment report, including the user's health status (such as heart rate stability and exercise adaptability).

[0139] Based on the assessment results, an exercise plan is recommended (such as 40 minutes of moderate-intensity running 3 times a week).

[0140] The evaluation results are presented to the user in a visual form through the user terminal display module.

[0141] The result: Users receive personalized health assessment reports and exercise recommendations.

[0142] Function: Improve the scientificity and adaptability of user experience and health management.

[0143] Step 5: Dynamic Feedback and Updates

[0144] Executing entities: user's wearable devices and cloud systems.

[0145] Trigger condition: Continuous upload of user daily data.

[0146] Action Description:

[0147] The system dynamically adjusts the user portrait and classification model results based on the newly added data.

[0148] Assessment reports and exercise recommendations are also optimized as new data is added.

[0149] Result: The system is updated in real time to ensure that the evaluation results always meet the user's latest status.

[0150] Function: To realize the dynamic adaptability of the system and improve the real-time and accuracy of health management.

[0151] Example 3

[0152] Adopt the above system to expand the technical solution

[0153] Technical solution: Application scenarios of group health management

[0154] In enterprise or community health management, the technical solution of the present invention can be expanded to a group health management system. And the following specific implementation steps can be performed:

[0155] Step 1: Data clustering collection

[0156] Execution entity: Wearable devices of corporate employees or community members.

[0157] Trigger condition: The system connects to multiple user terminals and starts synchronizing data.

[0158] Action Description:

[0159] Each member's exercise and health data is uploaded to the cloud platform through the terminal.

[0160] The data is processed centrally to generate a summary of the health characteristics of the group (such as exercise frequency, group average health score).

[0161] The result: a comprehensive dataset that reflects the overall health of the group.

[0162] Function: Support subsequent health management and intervention measures for the group.

[0163] Step 2: Group classification and health risk warning

[0164] Execution entity: cloud system.

[0165] Trigger condition: Complete the summary of group characteristics.

[0166] Action Description:

[0167] The system classifies the entire group (such as high health risk group, medium health risk group) based on the feature vector of each member.

[0168] Generate health warning reports for high-risk groups and recommend collective exercise intervention plans.

[0169] The results: health risk distribution maps and corresponding group exercise intervention plans.

[0170] Role: Helps managers understand the health of their group and develop collective optimization strategies.

[0171] By expanding the technical solutions of Example 2 and Example 3, the system of the present invention can flexibly adapt to personal health management and group health management to achieve accurate, personalized and efficient health assessment results.

[0172] At the same time, personalized sports health assessment based on user portraits and classification models can be realized, and the following effects can be achieved:

[0173] Effect 1: Improved accuracy of personalized health assessment

[0174] Through multi-source data collection and user portrait modeling, the system can integrate the user's physical characteristics, exercise habits and health status to achieve a multi-dimensional comprehensive assessment.

[0175] The introduction of the classification model refines the user groups, ensures that health assessment results and exercise recommendations are more accurate, and solves the problem of lack of deep personalization in existing technologies.

[0176] Effect 2: Enhanced system dynamic adaptability

[0177] The dynamic feedback and update mechanism ensures that the system can adjust user portraits and evaluation results in real time.

[0178] When the user's physical condition or exercise habits change, the evaluation results will be updated accordingly to keep the evaluation real-time and effective.

[0179] Effect 3: Improved user experience and trust

[0180] The evaluation report generated by the system is presented in a visual form, providing intuitive explanations and suggestions, which improves users' understanding and acceptance of the evaluation results.

[0181] Strongly explainable evaluation recommendations enhance users’ trust in the system and promote long-term use of health management.

[0182] Effect 4: Improved evaluation efficiency and scalability

[0183] Relying on cloud servers and automated algorithms, the system achieves efficient data processing and real-time evaluation.

[0184] The system supports large-scale user access and is suitable for personal health management and group health management scenarios such as enterprises / communities.

[0185] Expanded technical solutions using the above system: Application of group health management

[0186] Effect 21: Effective identification of group health risks

[0187] The group classification function can quickly identify people with high health risks, provide health warnings, and help companies or communities implement precise interventions.

[0188] Through visual analysis of overall health status, managers can intuitively understand group health trends and develop more targeted health management strategies.

[0189] Effect 22: Improved scientificity and efficiency of collective intervention

[0190] Based on the classification results, the system generates targeted exercise plans for different groups to improve the scientific nature and effectiveness of group health management.

[0191] The ability to centrally process data enables the system to support a large number of users while maintaining evaluation efficiency and reducing management costs.

[0192] Effect 23: Increased social benefits of health promotion

[0193] The implementation of a group health management system can effectively reduce the incidence of health problems among high-risk groups, indirectly reduce the consumption of medical resources, and improve the overall health level of society.

[0194] The above system can be used to derive recommendations with complete accuracy and personalization: through the combination of user portraits and classification models, the system can accurately analyze the user's health status and provide highly personalized exercise recommendations.

[0195] Dynamic adaptability: Supports real-time data updates to ensure that health assessment results always match the user's latest status.

[0196] Wide range of applications: The system design is compatible with individual and group health management needs and can meet a variety of application scenarios.

[0197] Efficiency and scalability: With the help of cloud platforms and automated algorithms, the system can efficiently process massive amounts of data and support large-scale user access.

[0198] Improved user experience: Through intuitive evaluation reports and explanatory suggestions, user trust and satisfaction are improved.

[0199] Enhanced social benefits: In group health management, the present invention helps to reduce health risks, improve health levels, and generate positive social benefits.

[0200] The above beneficial effects indicate that the present invention has significant improvements and advantages over the prior art in terms of technical implementation and practical application.

[0201] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A personalized sports assessment system, characterized in that: The system is based on the introduction of user portrait technology, accurately models the multi-dimensional characteristics of users, supports dynamic updates, uses classification models to achieve accurate user grouping, and provides users with personalized sports health assessment and recommendation reports based on classification results and user portraits. The system includes a control center module, and the control center module is connected to a data acquisition and preprocessing module, a user portrait generation module, a classification model operation and user grouping module, and a personalized health assessment and recommendation module. Each module is controlled and connected to the control center module respectively. The data acquisition module, user portrait generation module, classification model operation and user grouping module, and personalized health assessment and recommendation module are associated in sequence and transmit related information to each other.

2. A personalized sports assessment system according to claim 1, characterized in that: The data acquisition and preprocessing module is connected to the wearable device and the user terminal device, and collects user data in real time through the Bluetooth or Wi-Fi protocol, and completes the cleaning and normalization operations through the preprocessing module. The normalization formula is: Among them, x is the original data and x′ is the normalized data.

3. A personalized motion assessment system according to claim 2, characterized in that: The user portrait generation module extracts the static and dynamic features of the user through feature engineering based on the results of the data collection and preprocessing module, uses feature vectors to represent the user portrait, and maps the user features to a high-dimensional space. The feature extraction is: 1) Average heart rate: 2) Exercise frequency: count the number of days per week for exercise; 3) Feature vector representation: v = [age, gender, BMI, exercise frequency, average heart rate, ...] 4. A personalized motion assessment system according to claim 3, characterized in that: The classification model operation and user grouping module are connected to a support vector machine SVM, which uses a machine learning algorithm to classify user portraits to form a model, and is trained through historical annotation data to generate classification rules. The classification function algorithm of the support vector machine is: f(x)=sign(w·x+b) Among them, w is the weight vector, x is the user feature vector, and b is the bias.

5. A personalized motion assessment system according to claim 4, characterized in that: The personalized health assessment and recommendation module is designed with a health assessment formula: Health score = w1·BMI+w2·heart rate stability+w3·exercise frequency Generate a personalized health assessment report based on user grouping results, historical data and evaluation formulas, and recommend an exercise plan suitable for the user, including exercise type, intensity and frequency. Generate exercise plan matching: recommend appropriate exercise intensity based on the user's health score.

6. A personalized motion assessment system according to claim 2, characterized in that: The data acquisition and preprocessing module design value input code is:

7. The personalized sports assessment system according to claim 3, characterized in that: The user portrait generation module is designed based on the data acquisition and preprocessing module and the input code is:

8. The personalized sports assessment system according to claim 4, characterized in that: The classification model operation and user grouping module design input code is: # Example training data X_train = [[22.5,1,72],[25.0,0,85],[30.0,1,65]] # [BMI, gender, average heart rate] y_train = [0,1,0] #0: low intensity group, 1: high intensity group #Model training svm_model=SVC(kernel='linear') svm_model.fit(X_train,y_train) #User group prediction user_vector = [[23.0,1,78]] group=svm_model.predict(user_vector) print(f"User grouping results:{group[0]}").

9. The personalized sports assessment system according to claim 5, characterized in that: The personalized health assessment and recommendation module design input code is:

Images