Diet and exercise recommendation method and system based on digital twinborn

Abstract

The invention provides a diet and exercise recommendation method and system based on digital twin. The diet and exercise recommendation method based on digital twinning comprises the following steps: constructing a digital twinning model by utilizing collected initial physiological data and dynamic physiological data of a user; performing simulation prediction on the initial physiological data and the dynamic physiological data of the registered target user by using a digital twin model, and obtaining diet and operation types matched with the initial physiological data and the dynamic physiological data of the registered target user; and according to the diet and operation type matched with the initial physiological data and the dynamic physiological data of the registered target user, calling a diet and exercise scheme from a database, and recommending the diet and exercise scheme to the registered target user. The system comprises modules corresponding to the steps of the method.

Description

Technical Field

[0001] The present invention proposes a diet and exercise recommendation method and system based on digital twins, which belongs to the technical field of exercise and diet data recommendation. Background Art

[0002] As the pace of contemporary society accelerates and the public pays more attention to health, personalized health management plans are becoming a key factor in meeting people's health needs. Traditionally, diet and exercise recommendations often rely on universal health manuals or general guidance based on experience, which fails to fully consider the immediate impact of each person's unique physiological structure, living habits and daily environment on their health status.

[0003] Specifically, each person's physical condition, metabolic rate, genetic background, and daily activity level are different, and the above factors are crucial to developing an effective health management plan. At the same time, changes in the external environment, such as seasonal changes, work pressure, sleep quality, and even psychological state, will also have an impact on personal health that cannot be ignored. However, traditional recommendation methods often ignore the above-mentioned personalized nuances and dynamic changes, resulting in a significant reduction in the pertinence and effectiveness of the recommendations. Therefore, it is particularly important to develop a diet and exercise recommendation system that can accurately match individual needs and dynamically adjust recommendations. Summary of the invention

[0004] The present invention provides a diet and exercise recommendation method and system based on digital twins to solve the problems existing in the above-mentioned prior art. The technical solutions adopted are as follows: A diet and exercise recommendation method based on digital twins, the diet and exercise recommendation method based on digital twins comprising: Build a digital twin model using the collected initial physiological data and dynamic physiological data of the user; After the digital twin model is built, the initial physiological data of the registered target user is collected, and the dynamic physiological data of the target user is dynamically adjusted and collected according to the dynamic change state of the dynamic physiological data of the target user, so as to obtain the initial physiological data and dynamic physiological data of the registered target user; The digital twin model is used to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user.

[0005] Furthermore, the collected initial physiological data and dynamic physiological data of the user are used to construct a digital twin model, including: Retrieving the collected initial physiological data and dynamic physiological data of the user from the database; wherein the initial physiological data includes age, gender, weight, height and basal metabolism; the dynamic physiological data includes heart rate, blood pressure and exercise calories consumed; Cleaning the collected initial physiological data and dynamic physiological data of the user to remove invalid data, duplicate data and erroneous data, and obtaining the pre-processed initial physiological data and dynamic physiological data; A digital twin model is constructed using the preprocessed initial physiological data and dynamic physiological data, and the constructed digital twin model is trained and verified to obtain a trained and verified digital twin model.

[0006] Furthermore, the structure of the digital twin model is as follows: An input layer, used to input the initial physiological data and dynamic physiological data of the target user, and perform unified standardization processing on the initial physiological data and dynamic physiological data of the target user; A multi-scale time window layer, including multiple parallel LSTM sub-networks, and each LSTM sub-network corresponds to a time window; wherein the parallel LSTM sub-networks are used to process input sequences of different lengths to control the digital twin model to capture data patterns at different time scales; The adaptive gated fusion layer is used to dynamically adjust the parameter values ​​corresponding to the importance and contribution of each LSTM sub-network according to the current input initial physiological data and dynamic physiological data, and synthesize the output of the dynamically adjusted LSTM sub-network to form a unified hidden layer; A context-aware fusion layer that uses contextual information from the target user to combine with the output of the adaptive gating fusion layer; this dynamically adjusts the model’s response to make the output more personalized and precise; Output layer, used to output diet and running types.

[0007] Furthermore, the steps of obtaining the time window are as follows: Retrieve the preset basic decay rate from the database; Real-time monitoring of the frequency of data changes of the collected dynamic physiological data; The dynamic attenuation factor corresponding to each LSTM sub-network is obtained by using the preset basic attenuation rate and the data change frequency of the dynamic physiological data; The dynamic attenuation factor is used to obtain the time length of the time window corresponding to each LSTM sub-network; wherein the time length of the time window corresponding to each LSTM sub-network is obtained by the following formula: Among them, K irepresents the time length of the time window corresponding to the i-th LSTM sub-network; T represents the preset observation period length; α(i, β, λ) represents the dynamic attenuation factor corresponding to the i-th LSTM sub-network; N represents the total number of LSTM sub-networks; w i represents the preset initial weight value corresponding to the i-th LSTM sub-network, α0 represents the preset basic attenuation rate; β represents the change frequency factor of dynamic physiological data; i represents the index corresponding to each LSTM sub-network; λ represents the preset periodic intensity value of dynamic physiological data.

[0008] Furthermore, the dynamic attenuation factor corresponding to each LSTM subnetwork is obtained by using the preset basic attenuation rate and the data change frequency of the dynamic physiological data, including: Retrieve the data change frequency of the collected dynamic physiological data; The change frequency factor is obtained by using the data change frequency of the dynamic physiological data; wherein the change frequency factor is obtained by the following formula: Wherein, β represents the frequency factor of the dynamic physiological data; r represents the adjustment parameter, and the range of the adjustment parameter is [0.1, 0.9]; f 01 、f 02 and f 03 They correspond to the change ratios of the maximum changes in heart rate, blood pressure and exercise calories consumed; v p01 、v p02 and v p03 They correspond to the average rate of change of heart rate, blood pressure and exercise calories consumed per unit time; The dynamic attenuation factor corresponding to each LSTM sub-network is obtained by combining the change frequency factor with the basic attenuation rate; The dynamic attenuation factor corresponding to each LSTM sub-network is obtained by the following formula: Among them, α(i, β, λ) represents the dynamic attenuation factor corresponding to the i-th LSTM sub-network; α0 represents the preset basic attenuation rate; β represents the change frequency factor of the dynamic physiological data; i represents the index corresponding to each LSTM sub-network, and i=1,2,3,…,N; N represents the total number of LSTM sub-networks; λ represents the preset periodic intensity value of the dynamic physiological data.

[0009] Furthermore, after the digital twin model is built, the initial physiological data of the registered target user is collected, and the dynamic physiological data of the target user is dynamically adjusted and collected according to the dynamic change state of the dynamic physiological data of the target user, and the initial physiological data and dynamic physiological data of the registered target user are obtained, including: After the digital twin model is built, an initial physiological data input instruction is sent to the registered target user, and the initial physiological data input by the registered target user is received in real time; wherein the initial physiological data includes age, gender, weight, height and basal metabolism; The dynamic physiological data of the registered target user is collected in real time through a wearable device worn by the registered target user; wherein the dynamic physiological data includes heart rate, blood pressure and exercise calories consumed; The change frequency of the dynamic physiological data of the registered target user is extracted to dynamically adjust the collection frequency of the dynamic physiological data.

[0010] Further, extracting the change frequency of the dynamic physiological data of the registered target user to dynamically adjust the collection frequency of the dynamic physiological data includes: Extracting the change frequency of heart rate, blood pressure and exercise calorie consumption contained in the dynamic physiological data of the registered target user; Using the change frequencies of the heart rate, blood pressure and exercise calories consumed, obtaining the change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed; The acquisition frequency of the dynamic physiological data is adjusted using the change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed, wherein the adjusted acquisition frequency of the dynamic physiological data is obtained by the following formula: Among them, f new represents the adjusted acquisition frequency of dynamic physiological data; f y represents the acquisition frequency of dynamic physiological data before adjustment; ξ represents the preset adjustment coefficient, with a value range of 0.26-0.41; X 01 Indicates the frequency parameter corresponding to the heart rate; X 02 Indicates the frequency parameter of blood pressure change; X 03 Indicates the frequency of change corresponding to the calories consumed during exercise.

[0011] Further, using the change frequencies of the heart rate, blood pressure and exercise calories consumed to obtain the change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed, including: The frequency of change of the heart rate included in the dynamic physiological data is used to obtain the frequency of change parameter corresponding to the heart rate; wherein the frequency of change parameter corresponding to the heart rate is obtained by the following formula: Among them, X 01 Indicates the frequency parameter corresponding to the heart rate; X b01 Indicates the standard deviation of heart rate; m 01 Indicates the number of heart rate collections; H i represents the heart rate value corresponding to the i-th acquisition; H i+1 Indicates the heart rate value corresponding to the i+1th acquisition; H f Indicates the maximum allowable fluctuation range of heart rate value; The frequency of change of the blood pressure contained in the dynamic physiological data is used to obtain the frequency parameter of change corresponding to the blood pressure; wherein the frequency parameter of change corresponding to the blood pressure is obtained by the following formula: Among them, X 02 Indicates the frequency parameter of blood pressure change; X x02 represents the standard deviation of diastolic pressure; X y02 Indicates the standard deviation of systolic blood pressure; m 02 Indicates the number of blood pressure collections; S 01i+1 and S 01i represents the value corresponding to the diastolic pressure collected at the i+1th and ith times; S 02i+1 and S 02i Indicates the corresponding values ​​of systolic blood pressure collected at the i+1th and ith times; The frequency of change of the exercise calorie consumption contained in the dynamic physiological data is used to obtain the frequency parameter of change corresponding to the exercise calorie consumption; wherein the frequency parameter of change corresponding to the exercise calorie consumption is obtained by the following formula: Among them, X 03 Indicates the frequency parameter corresponding to the amount of calories consumed during exercise; X b03 Indicates the standard deviation of calories burned during exercise; m 03 Indicates the number of times the calories consumed during exercise are collected; E i Indicates the exercise calorie consumption value corresponding to the i-th collection; E i+1 Indicates the exercise calorie consumption value corresponding to the i+1th collection; E f Indicates the preset floating reference value for exercise calorie consumption.

[0012] Furthermore, the digital twin model is used to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user, including: Using the digital twin model to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, and obtaining a diet and running type that matches the initial physiological data and dynamic physiological data of the registered target user; According to the diet and running type matching the initial physiological data and dynamic physiological data of the registered target user, a diet and exercise plan is retrieved from the database and recommended to the registered target user.

[0013] A diet and exercise recommendation system based on digital twins, wherein the diet and exercise recommendation system based on digital twins comprises: A digital twin model building module, used to build a digital twin model using the collected initial physiological data and dynamic physiological data of the user; The data acquisition module is used to collect the initial physiological data of the registered target user after the digital twin model is built, and dynamically adjust and collect the dynamic physiological data of the target user according to the dynamic change state of the dynamic physiological data of the target user, so as to obtain the initial physiological data and dynamic physiological data of the registered target user; The plan recommendation module is used to use the digital twin model to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user.

[0014] Beneficial effects of the present invention: The digital twin model constructed by the diet and exercise recommendation method and system based on digital twin proposed in the present invention based on the initial physiological data and dynamic physiological data of the user can accurately reflect the physiological characteristics and change trends of each user, and provide a highly personalized basis for subsequent health management applications. Through steps such as data cleaning and model training verification, the accuracy and reliability of the digital twin model are effectively improved. The model can more accurately simulate and predict the physiological state and its dynamic changes of the user, and provide the user with more scientific health management suggestions. The digital twin model can receive and process the dynamic physiological data of the user in real time, and dynamically adjust the health management suggestions according to the changes in the user's physiological state. The above dynamic adaptability makes the health management service more in line with the actual needs of the user, and improves the pertinence and effectiveness of the service. The health management application based on the digital twin model can provide powerful decision-making support for professionals such as doctors and health managers. Through model simulation and prediction results, professionals can more accurately evaluate the health status of users, predict potential risks and formulate targeted health management plans. Personalized health management services and dynamic adaptive suggestions can significantly improve the user experience. Users can obtain tailored health management plans according to their actual conditions, better manage their health status, and improve their quality of life. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a flow chart of the method of the present invention; Figure 2 The system block diagram of the system of the present invention. DETAILED DESCRIPTION

[0016] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0017] The embodiment of the present invention proposes a diet and exercise recommendation method based on digital twins, such as Figure 1 As shown, the diet and exercise recommendation method based on digital twins includes: S1. Build a digital twin model using the collected initial physiological data and dynamic physiological data of the user; S2. After the digital twin model is built, the initial physiological data of the registered target user is collected, and the dynamic physiological data of the target user is dynamically adjusted and collected according to the dynamic change state of the dynamic physiological data of the target user, so as to obtain the initial physiological data and dynamic physiological data of the registered target user; S3. Use the digital twin model to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user.

[0018] The working principle of the above technical solution is: retrieve the user's historical physiological data from the database, including static initial physiological data (such as age, gender, weight, height and basal metabolism) and dynamic physiological data (such as heart rate, blood pressure and exercise calories consumed). The above data is the basis for building a digital twin model. Cleaning the retrieved data is an important data quality control process. By removing invalid data (such as data with too many missing values ​​or data that is obviously inconsistent with common sense), duplicate data and erroneous data, the accuracy and reliability of subsequent model construction are ensured. Use the preprocessed data to build a digital twin model. Train and verify the constructed digital twin model. The training process is to adjust the model parameters through a large amount of data so that the model can more accurately reflect the actual situation; the verification process is to evaluate the performance of the model on unseen data to ensure the generalization ability and stability of the model.

[0019] The effect of the above technical solution is: the digital twin model constructed based on the user's initial physiological data and dynamic physiological data can accurately reflect the physiological characteristics and changing trends of each user, and provide a highly personalized basis for subsequent diet and exercise recommendations. Through data cleaning and model training verification, the accuracy and reliability of the digital twin model are effectively improved, enabling it to more accurately predict the user's physiological response under different diet and exercise plans. The model can receive and process the user's dynamic physiological data in real time, and dynamically adjust the diet and exercise recommendation plan according to changes in the user's physiological state to ensure the scientificity and timeliness of the recommendations. The diet and exercise recommendation method based on the digital twin model can provide users with more accurate and effective health management services, help users better manage their health status, prevent the occurrence of diseases, and improve the quality of life.

[0020] In summary, the above technical solution achieves accurate management and effective intervention of users' health status by building a highly personalized digital twin model and combining dynamic physiological data to update recommended solutions in real time.

[0021] In one embodiment of the present invention, a digital twin model is constructed using the collected initial physiological data and dynamic physiological data of the user, including: S101, retrieving the collected initial physiological data and dynamic physiological data of the user from the database; wherein the initial physiological data includes age, gender, weight, height and basal metabolism; the dynamic physiological data includes heart rate, blood pressure and exercise calories consumed; S102, performing data cleaning on the collected initial physiological data and dynamic physiological data of the user, removing invalid data, duplicate data and erroneous data, and obtaining pre-processed initial physiological data and dynamic physiological data; S103: construct a digital twin model using the preprocessed initial physiological data and dynamic physiological data, and train and verify the constructed digital twin model to obtain a trained and verified digital twin model.

[0022] The working principle of the above technical solution is: retrieve the user's initial physiological data and dynamic physiological data from a pre-established database. The above data is the basis for building a digital twin model, where the initial physiological data includes static information that is not easy to change frequently, such as age, gender, weight, height and basal metabolic rate; dynamic physiological data includes real-time changing information, such as heart rate, blood pressure and exercise calories consumed. Data cleaning is performed on the collected raw data, which is a key step to ensure the accuracy of the model. Data cleaning includes removing invalid data (such as too many missing values ​​and obviously abnormal data points), duplicate data (completely identical data generated due to system errors or repeated entries) and erroneous data (such as data that obviously does not conform to physiological common sense). Through the above technical solution, clean, accurate and reliable pre-processed data can be obtained to provide a high-quality data source for subsequent model construction. Using the pre-processed data, combined with the principles and methods of digital twin technology, a digital twin model that can reflect the user's physiological state and its dynamic changes is constructed. This process involves selecting appropriate mathematical models, algorithms and parameters to achieve accurate simulation and prediction of the user's physiological characteristics. The constructed digital twin model is trained and verified. The training process is to adjust the model parameters through a large amount of historical data so that the model can more accurately reflect the user's physiological state and its changing trends; the verification process is to evaluate the performance of the model on unseen data to ensure that the model has good generalization ability and stability. Through the above technical solutions, a digital twin model that has been trained and verified can be obtained, providing strong support for subsequent health management applications.

[0023] The effect of the above technical solution is: the digital twin model constructed based on the user's initial physiological data and dynamic physiological data can accurately reflect the physiological characteristics and change trends of each user, and provide a highly personalized foundation for subsequent health management applications. Through steps such as data cleaning and model training verification, the accuracy and reliability of the digital twin model are effectively improved. The model can more accurately simulate and predict the user's physiological state and its dynamic changes, and provide users with more scientific health management suggestions. The digital twin model can receive and process the user's dynamic physiological data in real time, and dynamically adjust the health management suggestions according to the changes in the user's physiological state. The above dynamic adaptability makes the health management service more in line with the actual needs of the user, and improves the pertinence and effectiveness of the service. The health management application based on the digital twin model can provide strong decision-making support for professionals such as doctors and health managers. Through model simulation and prediction results, professionals can more accurately assess the user's health status, predict potential risks and formulate targeted health management plans. Personalized health management services and dynamic adaptive suggestions can significantly improve the user experience. Users can obtain tailored health management plans according to their actual conditions, better manage their health status, and improve their quality of life.

[0024] In one embodiment of the present invention, the structure of the digital twin model is as follows: An input layer, used to input the initial physiological data and dynamic physiological data of the target user, and perform unified standardization processing on the initial physiological data and dynamic physiological data of the target user; A multi-scale time window layer, including multiple parallel LSTM sub-networks, and each LSTM sub-network corresponds to a time window; wherein the parallel LSTM sub-networks are used to process input sequences of different lengths to control the digital twin model to capture data patterns at different time scales; The adaptive gated fusion layer is used to dynamically adjust the parameter values ​​corresponding to the importance and contribution of each LSTM sub-network according to the current input initial physiological data and dynamic physiological data, and synthesize the output of the dynamically adjusted LSTM sub-network to form a unified hidden layer; A context-aware fusion layer that uses contextual information from the target user (e.g., current activity type, time, etc.) to combine with the output of the adaptive gating fusion layer; this dynamically adjusts the model’s response to make the output more personalized and precise; Output layer, used to output diet and running types.

[0025] The working principle of the above technical solution is as follows: the input layer is responsible for receiving the initial physiological data and dynamic physiological data of the target user. The above data is the basis for model prediction, including static initial physiological data (such as age, gender, weight, etc.) and real-time dynamic physiological data (such as heart rate, blood pressure, exercise calories consumed, etc.).

[0026] The input layer will also perform unified standardization on the above data to ensure that data from different sources and different magnitudes can be fairly compared in the model, thereby improving the model's prediction accuracy.

[0027] Multi-scale time window layer: This layer contains multiple parallel LSTM (Long Short-Term Memory) subnetworks, each corresponding to a specific time window. LSTM is a special recurrent neural network (RNN) that can effectively handle long-term dependencies in sequence data.

[0028] By setting up multiple parallel LSTM sub-networks, the model is able to simultaneously capture data patterns at different time scales, which is crucial for understanding the long-term trends and short-term fluctuations of the user's physiological state.

[0029] Adaptive Gating Fusion Layer: This layer dynamically adjusts the importance and contribution of each LSTM sub-network according to the current input initial physiological data and dynamic physiological data. This is achieved by learning a set of parameters that reflect the importance of information at different time scales for the current prediction task.

[0030] The adaptive gated fusion layer synthesizes the dynamically adjusted LSTM sub-network outputs to form a unified hidden layer representation. This representation integrates information at different time scales and provides a rich contextual basis for subsequent predictions.

[0031] Context-aware fusion layer: This layer uses contextual information from the target user (such as current activity type, time, etc.) to further adjust the model's response. The above information is crucial to understanding the user's current physiological state and needs.

[0032] By combining contextual information with the output of the adaptive gating fusion layer, the model can generate more personalized and accurate prediction results. The above combination enables the model to flexibly adjust the prediction strategy according to the actual situation of the user.

[0033] Output layer: The output layer is responsible for generating the final diet and exercise type recommendations. The above recommendations are based on the processing results of the previous layers, taking into account the user's initial physiological state, dynamic physiological changes, and contextual information.

[0034] The effect of the above technical solution is: by comprehensively considering the user's initial physiological data, dynamic physiological data and contextual information, the model can generate highly personalized diet and exercise type recommendations. The above recommendations are closer to the user's actual needs and living habits. Using the multi-scale time window layer, the model can simultaneously capture data patterns on different time scales. This enables the model to more comprehensively understand the user's physiological state change trends and short-term fluctuations, and improve the accuracy and robustness of the prediction. The adaptive gating fusion layer and the context-aware fusion layer enable the model to dynamically adjust the prediction strategy based on the user's real-time input and contextual information. The above adaptive adjustment capability enables the model to better adapt to complex and changing actual situations and improve the flexibility and accuracy of the prediction. Based on the above working principles and technical effects, the digital twin model can provide users with more accurate, personalized and practical diet and exercise type recommendations. This will help users better manage their health conditions, improve their quality of life and enhance user experience.

[0035] The above model uses a multi-scale time window layer to set up LSTM sub-networks with different time window sizes in parallel, so that each sub-network can focus on capturing data patterns of a specific time length. The above design allows the model to understand and integrate short-term, medium-term and long-term data changes at the same time. For example, LSTM with a short time window is more sensitive to rapid heart rate changes, while LSTM with a long time window can identify long-term physiological trends or periodic patterns. Through multi-scale time windows, the model can understand data from multiple dimensions, which not only increases the expressiveness of the data, but also helps the model to understand complex physiological and behavioral patterns more comprehensively. The above multi-level time series analysis is of great significance for predicting the health status and behavior of users, and can provide more accurate health advice. In practical applications, the physiological data of users will be affected by many factors and show different dynamics and complexity. The multi-scale time window layer enhances the adaptability of the model to different data characteristics by processing sequences of different lengths. For example, when daily activities are relatively calm, long-term patterns are more important; while when performing high-intensity exercise, short-term heart rate changes are more critical. After the multi-scale time window layer, the outputs of each sub-network are intelligently fused through an adaptive gating mechanism, which can dynamically adjust the contribution of each time scale output according to the current input. The above dynamic weight adjustment mechanism provides the model with more flexible data processing capabilities, enabling the model to respond optimally to different types of inputs. The multi-scale time window layer enables the model to learn and predict at multiple time scales, thereby better adapting to and processing complex and dynamically changing physiological data. The design of the above layers not only improves the performance of the model, but also enhances the applicability and flexibility of the model for different application scenarios.

[0036] In one embodiment of the present invention, the steps of acquiring the time window are as follows: Step 1, retrieve the preset basic decay rate from the database; Step 2: real-time monitoring of the data change frequency of the collected dynamic physiological data; Step 3: Obtain the dynamic attenuation factor corresponding to each LSTM subnetwork using the preset basic attenuation rate and the data change frequency of the dynamic physiological data; Step 4: Use the dynamic attenuation factor to obtain the time length of the time window corresponding to each LSTM subnetwork; wherein the time length of the time window corresponding to each LSTM subnetwork is obtained by the following formula: Among them, K irepresents the time length of the time window corresponding to the i-th LSTM sub-network; T represents the preset observation period length; α(i, β, λ) represents the dynamic attenuation factor corresponding to the i-th LSTM sub-network; N represents the total number of LSTM sub-networks; w i represents the preset initial weight value corresponding to the i-th LSTM sub-network, α0 represents the preset basic attenuation rate; β represents the change frequency factor of dynamic physiological data; i represents the index corresponding to each LSTM sub-network; λ represents the preset periodic intensity value of dynamic physiological data.

[0037] The working principle of the above technical solution is as follows: First, a preset basic decay rate is retrieved from the database. This decay rate is a basic parameter used for subsequent calculation of the dynamic decay factor, which reflects the general law of data changes over time.

[0038] Monitor the frequency of data changes of the collected dynamic physiological data in real time. This step is the key to obtaining the time window, because the frequency of changes in dynamic physiological data directly reflects the activity level and change trend of the user's physiological state. Using the preset basic decay rate and the frequency of data changes in dynamic physiological data, calculate the dynamic decay factor corresponding to each LSTM subnetwork. This factor reflects the decay speed of the data on the current time scale and is an important basis for determining the length of the time window.

[0039] According to the dynamic decay factor, the time length of the time window corresponding to each LSTM sub-network is calculated using the above formula. The above formula takes into account the observation period, the dynamic decay factor, the total number of LSTM sub-networks, and the preset initial weight value corresponding to each sub-network, ensuring that the length of the time window can reflect the dynamic changes of the data while maintaining the stability and prediction performance of the model.

[0040] The effect of the above technical solution is: by real-time monitoring of the frequency of changes in dynamic physiological data, and calculating the dynamic attenuation factor and time window length accordingly, the model can dynamically adapt to changes in the user's physiological state. This makes the model more flexible and accurate when processing data from different users and at different time points. Different time windows are assigned to each LSTM subnetwork, so that the model can capture data patterns on different time scales at the same time. It can more comprehensively understand the user's physiological state and improve the accuracy and robustness of the prediction. The dynamic adjustment of the time window takes into account the unique physiological characteristics of each user and the changing characteristics of dynamic physiological data. This enables the model to provide more personalized predictions and recommendations for each user and improve the user experience. By reasonably setting the length of the time window, the performance and stability of the LSTM subnetwork can be optimized. It can reduce the computational complexity of the model, improve the prediction speed, and maintain a high prediction accuracy.

[0041] By retrieving the preset basic decay rate from the database, the above technical solution provides a basic framework for determining the time window. This preset value can be adjusted according to different application scenarios and physiological data, thereby increasing the flexibility of the solution. The data change frequency of the collected dynamic physiological data is monitored in real time, so that the determination of the time window can dynamically adapt to the change of the data. It can ensure that the setting of the time window matches the actual situation of the data, and improve the accuracy of the analysis. At the same time, the dynamic decay factor corresponding to each LSTM subnetwork is calculated using the preset basic decay rate and the data change frequency of the dynamic physiological data. The above technical solution takes into account the dynamic characteristics of the data and the characteristics of the LSTM subnetwork. The introduction of the dynamic decay factor enables the length of the time window to be dynamically adjusted as the data changes, which helps to capture key information in the data. The time length of the time window corresponding to each LSTM subnetwork is calculated by the formula, and the above technical solution realizes the precise control of the length of the time window. This precise calculation helps to ensure that the changing trends and characteristics of the data can be accurately captured during the analysis process.

[0042] The above formula takes into account multiple factors, including the preset observation period length, dynamic decay factor, the total number of LSTM subnetworks, and the preset initial weight value corresponding to each LSTM subnetwork. The combined effect of these factors makes the length of the time window more consistent with the actual situation of the data and analysis requirements. By setting different time window lengths for different LSTM subnetworks, the above technical solution can more accurately capture the data features of each subnetwork. It can improve analysis efficiency and accuracy, making the analysis results more reliable.

[0043] At the same time, the above technical solution also takes into account the dynamic characteristics of the data and the characteristics of the LSTM sub-network, so that it can better adapt to different types of data and analysis needs. By clarifying the acquisition steps and calculation methods of the time window, the above technical solution enhances the interpretability of the model. This allows researchers to have a clearer understanding of how the model processes data and makes predictions, which helps to further improve and optimize the model.

[0044] In summary, the above technical solution enables the digital twin model to better adapt to changes in the user's physiological state by dynamically adjusting the length of the time window, improving the accuracy and personalization of predictions, and thus providing users with more accurate and practical health management suggestions. At the same time, the above technical solution achieves flexible and precise control of the time window by introducing dynamic attenuation factors and accurately calculating the length of the time window. In addition, it can improve analysis efficiency and accuracy, enhance the interpretability of the model, and provide better support for different types of data and analysis needs.

[0045] In one embodiment of the present invention, the dynamic attenuation factor corresponding to each LSTM sub-network is obtained by using the preset basic attenuation rate and the data change frequency of the dynamic physiological data, including: Step 301, retrieve the data change frequency of the collected dynamic physiological data; Step 302: Obtain a change frequency factor using the data change frequency of the dynamic physiological data; wherein the change frequency factor is obtained by the following formula: Wherein, β represents the frequency factor of the dynamic physiological data; r represents the adjustment parameter, and the range of the adjustment parameter is [0.1, 0.9]; f 01 、f 02 and f 03 They correspond to the change ratios of the maximum changes in heart rate, blood pressure and exercise calories consumed; v p01 、v p02 and v p03 They correspond to the average rate of change of heart rate, blood pressure and exercise calories consumed per unit time; Step 303: Obtain a dynamic attenuation factor corresponding to each LSTM subnetwork by using the change frequency factor combined with the basic attenuation rate; The dynamic attenuation factor corresponding to each LSTM sub-network is obtained by the following formula: Among them, α(i, β, λ) represents the dynamic attenuation factor corresponding to the i-th LSTM sub-network; α0 represents the preset basic attenuation rate; β represents the change frequency factor of the dynamic physiological data; i represents the index corresponding to each LSTM sub-network, and i=1,2,3,…,N; N represents the total number of LSTM sub-networks; λ represents the preset periodic intensity value of the dynamic physiological data.

[0046] The working principle of the above technical solution is as follows: first, retrieve the data change frequency of the collected dynamic physiological data. The above data change frequency reflects the activity level of the user's physiological state at different time points. The data change frequency of the dynamic physiological data is combined with the above formula to calculate the change frequency factor. The above formula takes into account the change ratio corresponding to the maximum change amplitude of heart rate, blood pressure and exercise calories consumed, as well as their average change rate per unit time. At the same time, an adjustment parameter is introduced to adjust the sensitivity of the change frequency factor. Combining the change frequency factor and the basic decay rate, another formula is used to calculate the dynamic decay factor corresponding to each LSTM subnetwork. The above formula takes into account the index of the LSTM subnetwork, the basic decay rate, the change frequency factor, and the preset periodic intensity value of the dynamic physiological data. Through the above formula, a suitable dynamic decay factor can be assigned to each LSTM subnetwork for subsequent time window length calculation.

[0047] The effect of the above technical solution is: by real-time monitoring of the frequency of changes in dynamic physiological data and calculating the dynamic attenuation factor based on it, the model can adapt to changes in the user's physiological state more flexibly. This enables the model to more accurately capture the changing trend of the data when processing data from different users and different time points. Assigning different dynamic attenuation factors to each LSTM subnetwork can make the model more accurate in capturing data patterns on different time scales. It can more comprehensively understand the user's physiological state and improve the accuracy and robustness of the prediction. The calculation of the dynamic attenuation factor takes into account the unique physiological characteristics of each user and the changing characteristics of dynamic physiological data. This enables the model to provide more personalized predictions and recommendations for each user to better meet the actual needs of the user. By reasonably setting the dynamic attenuation factor, the performance and stability of the LSTM subnetwork can be optimized. It can reduce the computational complexity of the model, improve the prediction speed, and maintain a high prediction accuracy.

[0048] In summary, the above technical solution dynamically adjusts the dynamic attenuation factor corresponding to each LSTM sub-network, so that the digital twin model can better adapt to changes in the user's physiological state, improve the accuracy and personalization of prediction, and thus provide users with more accurate and practical health management suggestions.

[0049] In one embodiment of the present invention, after the digital twin model is constructed, initial physiological data of a registered target user is collected, and the dynamic physiological data of the target user is dynamically adjusted and collected according to the dynamic change state of the dynamic physiological data of the target user, so as to obtain the initial physiological data and dynamic physiological data of the registered target user, including: S201. After the digital twin model is built, an initial physiological data input instruction is sent to a registered target user, and the initial physiological data input by the registered target user is received in real time; wherein the initial physiological data includes age, gender, weight, height and basal metabolism; S202, collecting dynamic physiological data of the registered target user in real time through a wearable device worn by the registered target user; wherein the dynamic physiological data includes heart rate, blood pressure and exercise calories consumed; S203: extracting the change frequency of the dynamic physiological data of the registered target user to dynamically adjust the collection frequency of the dynamic physiological data.

[0050] The working principle of the above technical solution is as follows: after the digital twin model is built, the system first sends an initial physiological data input instruction to the registered target user. The above instruction guides the user to enter his basic physiological information, such as age, gender, weight, height and basal metabolic rate. The above information is the basic static data for establishing the user's digital twin model. After the user enters the above data through the system interface or designated channels, the system receives and stores the above initial physiological data in real time. The system collects dynamic physiological data in real time through the wearable device worn by the registered user. The above wearable device can continuously monitor the user's key physiological indicators such as heart rate, blood pressure, exercise calories consumed, and transmit the above data to the system in real time. During the collection process, the system extracts the frequency of change of dynamic physiological data, that is, the speed of change of the above physiological indicators over time. According to the frequency of change, the system dynamically adjusts the collection frequency to ensure that data can be collected more frequently when the data changes faster to capture more detailed change trends; when the data changes slowly, the collection frequency can be appropriately reduced to save resources.

[0051] The effect of the above technical solution is: by collecting the user's initial physiological data and dynamic physiological data, the system can build a highly personalized digital twin model. The above model can more accurately reflect the user's physiological state and change trend, providing a solid foundation for subsequent health management and prediction. Real-time collection of dynamic physiological data and dynamic adjustment of the collection frequency according to the frequency of data changes ensure that the system can capture subtle changes in the user's physiological state and improve the accuracy of prediction.

[0052] Wearable devices collect data in real time, allowing the system to instantly obtain the user's latest physiological status information. The above real-time performance provides the system with the ability to respond dynamically, and can issue warnings or suggestions in time when the user's physiological state is abnormal. The mechanism of dynamically adjusting the acquisition frequency further enhances the system's real-time response capability, ensuring that sufficient data can be obtained to support decision-making at critical moments. By dynamically adjusting the acquisition frequency according to the frequency of data changes, the system can reasonably optimize resource usage while ensuring data accuracy. Reducing the acquisition frequency when data changes slowly can reduce the load of data transmission and storage, and reduce system energy consumption and costs. Users only need to wear wearable devices and enter a small amount of initial physiological data to enjoy personalized, real-time health management services. The above convenient operation method improves the user experience and makes users more willing to participate in the health management process.

[0053] In summary, the above technical solutions provide strong support for the application of digital twin models in the field of health management through efficient and accurate physiological data collection mechanisms. It not only improves the accuracy and real-time performance of predictions, but also optimizes resource usage and improves user experience.

[0054] In one embodiment of the present invention, extracting the change frequency of the dynamic physiological data of the registered target user to dynamically adjust the collection frequency of the dynamic physiological data includes: S2031, extracting the change frequency of the heart rate, blood pressure and exercise calorie consumption contained in the dynamic physiological data of the registered target user; S2032, using the change frequencies of the heart rate, blood pressure and exercise calories consumed to obtain change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed; S2033, adjusting the acquisition frequency of the dynamic physiological data using the change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed, wherein the adjusted acquisition frequency of the dynamic physiological data is obtained by the following formula: Among them, f new represents the adjusted acquisition frequency of dynamic physiological data; f y represents the acquisition frequency of dynamic physiological data before adjustment; ξ represents the preset adjustment coefficient, with a value range of 0.26-0.41; X 01 Indicates the frequency parameter corresponding to the heart rate; X 02 Indicates the frequency parameter of blood pressure change; X 03 Indicates the frequency of change corresponding to the calories consumed during exercise.

[0055] The working principle of the above technical solution is as follows: the system first extracts the dynamic physiological data of the registered target user, which includes key physiological indicators such as heart rate, blood pressure and exercise calories consumed. Then, the system analyzes the time series data of the above physiological indicators and calculates their respective change frequencies. The above change frequencies reflect the speed of changes of physiological indicators over time.

[0056] The system uses the extracted change frequencies of heart rate, blood pressure and exercise calories consumed to calculate the corresponding change frequency parameters through a certain algorithm or formula. The above parameters are further processing or quantification of the original change frequency, which is used to adjust the subsequent acquisition frequency. Based on the calculated change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed, and the preset adjustment coefficient (such as ξ), the system calculates the adjusted dynamic physiological data acquisition frequency (fnew) through the above formula. The above formula comprehensively considers the original acquisition frequency (fy), the change frequency parameters of each physiological indicator (X01, X02, X03) and the preset adjustment coefficient (ξ), ensuring that the acquisition frequency can be adaptively adjusted according to the dynamic changes of physiological data.

[0057] The effect of the above technical solution is: by dynamically adjusting the collection frequency, the system can collect data more frequently when the physiological data changes rapidly, thereby more accurately capturing the changing trend of the physiological state. It can improve the accuracy of subsequent health prediction or diagnosis. When the physiological data changes slowly, the system appropriately reduces the collection frequency, reduces the load of data transmission and storage, and thus saves resources (such as power, network bandwidth, etc.). The above dynamic adjustment mechanism helps to extend the service life of wearable devices and reduce overall operating costs.

[0058] For users, dynamically adjusting the collection frequency means that in most cases, the system will not disturb them too frequently (such as through vibration or reminders). At the same time, when necessary (such as when the physiological state is abnormal), the system can collect data in a timely manner and issue warnings or suggestions, thereby improving user experience and satisfaction. The above technical solution provides a flexible data collection mechanism for the application of digital twin models in the field of health management. It allows the system to dynamically adjust the collection strategy according to the user's actual physiological state, so as to better adapt to the needs of different users and different scenarios.

[0059] In summary, the above technical solution achieves a dual improvement in the accuracy and efficiency of data collection by extracting the changing frequency of dynamic physiological data and adjusting the collection frequency accordingly, and has important practical application value.

[0060] In one embodiment of the present invention, the changing frequencies of the heart rate, blood pressure and exercise calories consumed are used to obtain the changing frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed, including: Step 1: Utilize the change frequency of the heart rate included in the dynamic physiological data to obtain the change frequency parameter corresponding to the heart rate; wherein the change frequency parameter corresponding to the heart rate is obtained by the following formula: Among them, X 01 Indicates the frequency parameter corresponding to the heart rate; X b01 Indicates the standard deviation of heart rate; m 01 Indicates the number of heart rate collections; H i represents the heart rate value corresponding to the i-th acquisition; H i+1 Indicates the heart rate value corresponding to the i+1th acquisition; H f Indicates the maximum allowable fluctuation range of heart rate value; Step 2: using the change frequency of the blood pressure contained in the dynamic physiological data to obtain the change frequency parameter corresponding to the blood pressure; wherein the change frequency parameter corresponding to the blood pressure is obtained by the following formula: Among them, X 02 Indicates the frequency parameter of blood pressure change; X x02 represents the standard deviation of diastolic pressure; X y02 Indicates the standard deviation of systolic blood pressure; m 02 Indicates the number of blood pressure collections; S 01i+1 and S 01i represents the value corresponding to the diastolic pressure collected at the i+1th and ith times; S 02i+1 and S 02i Indicates the corresponding values ​​of systolic blood pressure collected at the i+1th and ith times; Step 3: using the frequency of change of exercise calories consumed contained in the dynamic physiological data, obtain the frequency parameter corresponding to the exercise calories consumed; wherein the frequency parameter corresponding to the exercise calories consumed is obtained by the following formula: Among them, X 03 Indicates the frequency parameter corresponding to the amount of calories consumed during exercise; X b03 Indicates the standard deviation of calories burned during exercise; m 03 Indicates the number of times the calories consumed during exercise are collected; E i Indicates the exercise calorie consumption value corresponding to the i-th collection; E i+1 Indicates the exercise calorie consumption value corresponding to the i+1th collection; E f Indicates the preset floating reference value for exercise calorie consumption.

[0061] The working principle of the above technical solution is as follows: First, the system extracts the heart rate data from the dynamic physiological data and calculates the standard deviation of the heart rate (Xb01 ), to measure the discreteness of the heart rate data. Then, the system uses the number of heart rate data collection (m 01 ) and the heart rate values ​​collected twice consecutively (H i and H i+1 ), the change frequency parameter corresponding to the heart rate is calculated by the above formula (X 01 ). The above formula takes into account the overall discreteness of the heart rate data and the changes between adjacent data points, thus reflecting the frequency characteristics of the heart rate changes.

[0062] The system extracts blood pressure data from dynamic physiological data, including diastolic pressure and systolic pressure, and calculates their standard deviations (X x02 and X y02 Then, the system uses the number of blood pressure data collection times (m 02 ) and the diastolic and systolic blood pressure values ​​collected twice consecutively (S 01i , S 01i+1 , S 02i and S 02i+1 ), the frequency parameter corresponding to the blood pressure is calculated by the above formula (X 02 ). The above formula takes into account the changing characteristics of diastolic and systolic blood pressure, as well as the changes between adjacent data points, so that it can more accurately reflect the frequency of blood pressure changes. The system extracts the exercise calorie consumption data from the dynamic physiological data and calculates its standard deviation (X b03 ).

[0063] Next, the system uses the number of times the exercise calorie consumption data is collected (m 03 ) and the exercise calorie consumption values ​​collected twice in a row (E i and E i+1 ), the frequency parameter corresponding to the exercise calories consumed is calculated by the above formula (X 03 ). The above formula takes into account the overall discreteness of the exercise calorie consumption data and the changes between adjacent data points, thus reflecting the frequency characteristics of the changes in exercise calorie consumption.

[0064] The effect of the above technical solution is: by calculating the change frequency parameters corresponding to heart rate, blood pressure and exercise calories consumed, the system can more accurately reflect the changes in the user's physiological state. The above parameters comprehensively consider the overall discreteness of physiological data and the changes between adjacent data points, so as to provide more accurate and comprehensive physiological state information. The calculation of the change frequency parameters provides an important basis for the system to dynamically adjust the acquisition frequency of physiological data. When the physiological state changes rapidly, the system can capture more physiological information by increasing the acquisition frequency; when the physiological state changes slowly, the system can appropriately reduce the acquisition frequency to save resources. At the same time, accurate calculation of the change frequency parameters helps to improve the accuracy and efficiency of health management. The system can provide personalized health management suggestions or warnings based on the user's actual physiological state, so as to better protect the user's health. By dynamically adjusting the acquisition frequency and providing accurate physiological state information, the system can optimize the user experience. Users can experience more personalized and accurate health management services, thereby improving their satisfaction and trust in the system.

[0065] On the other hand, by calculating the frequency of changes corresponding to heart rate, blood pressure, and exercise calories consumed, the system can quickly identify significant changes in the user's physiological state. This helps the system respond more quickly, such as issuing an alarm or adjusting the health management strategy in a timely manner when an abnormal physiological state is detected. This real-time performance is essential for monitoring and intervention in emergency health conditions. The introduction of frequency of change parameters provides a new dimension for the analysis of physiological data. In addition to traditional statistical indicators such as mean and standard deviation, the system can also use these parameters to deeply analyze the dynamic changes of the user's physiological state. This helps to discover potential physiological problems or trends and provide more comprehensive and in-depth insights for health management. At the same time, by accurately calculating the frequency of changes in the user's heart rate, blood pressure, and exercise calories consumed, the system can more accurately understand the user's physiological characteristics and needs. This helps the system provide more personalized health management services, such as customized exercise plans and dietary recommendations based on the user's physiological state. This personalized service can better meet the needs of users and improve the effectiveness of health management.

[0066] By dynamically adjusting the frequency of physiological data collection, the system can reasonably allocate resources according to the user's actual physiological state. When the physiological state is stable, reducing the collection frequency can reduce unnecessary resource consumption; when the physiological state changes greatly, increasing the collection frequency can ensure that enough data is obtained to support accurate health analysis and decision-making. This resource allocation strategy helps to reduce the operating cost of the system and improve overall efficiency. When users see the accurate health advice and warnings provided by the system based on their changes in physiological state, they will pay more attention to their health status. This increase in attention helps to enhance users' health awareness and encourage them to participate more actively in health management. The active participation of users can not only improve the effectiveness of health management, but also enhance their trust and satisfaction with the system.

[0067] In summary, the above technical solution realizes the function of accurately capturing changes in the user's physiological state and dynamically adjusting the acquisition frequency by calculating the frequency parameters of changes corresponding to heart rate, blood pressure and exercise calories consumed, and has important practical application value. By calculating the frequency parameters of changes corresponding to heart rate, blood pressure and exercise calories consumed, this technical solution not only improves the accuracy and efficiency of health management, but also brings about technical effects in real time, personalized services, resource optimization, and user health awareness enhancement. These effects jointly promote the improvement of user health status and the improvement of health management level.

[0068] In one embodiment of the present invention, a digital twin model is used to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user, including: S301, using a digital twin model to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, and obtaining a diet and running type that matches the initial physiological data and dynamic physiological data of the registered target user; S302: According to the diet and exercise types that match the initial physiological data and dynamic physiological data of the registered target user, a diet and exercise plan is retrieved from a database and recommended to the registered target user.

[0069] The working principle of the above technical solution is as follows: First, the system collects and integrates the initial physiological data (such as age, gender, weight, height and basal metabolism) and dynamic physiological data (such as heart rate, blood pressure and exercise calories consumed) of the registered target users. Subsequently, the above physiological data are input into the pre-built digital twin model. The digital twin model uses advanced algorithms and machine learning techniques to simulate and predict the user's physiological state. By simulating physiological reactions and change trends under different physiological states, the model can evaluate the potential effects of users under different diets and exercise plans. Based on the results of simulation and prediction, the system can identify the diet and exercise type that best matches the user's current physiological state. For example, for users who need to lose weight, the model will recommend a low-fat, high-protein diet and a moderate-intensity aerobic exercise plan.

[0070] Once the diet and exercise types that match the user's physiological data are determined, the system will retrieve the corresponding diet and exercise plans from the preset database. The above plans include specific diet plans (such as recipes for three meals a day, food intake, etc.) and exercise plans (such as exercise type, exercise time, exercise intensity, etc.). Finally, the system recommends the retrieved personalized diet and exercise plans to the registered target users. The recommended content is presented in the form of text, graphics or application interface, which is convenient for users to understand and execute.

[0071] The effect of the above technical solution is: through the simulation and prediction capabilities of the digital twin model, the system can provide highly personalized diet and exercise plan recommendations for each user. The above personalized recommendations help users manage their health more accurately and improve their quality of life and life expectancy. The digital twin model can simulate and predict the changing trends of the user's physiological state in real time, thereby providing a scientific basis for the formulation of diet and exercise plans. The above accurate prediction helps to optimize the effect of the plan and reduce unnecessary attempts and adjustments. Personalized diet and exercise plan recommendations can significantly improve the user experience and satisfaction. Users no longer need to blindly try various diet and exercise plans, but can directly obtain the health management advice that best suits them.

[0072] Through continuous personalized recommendations and feedback mechanisms, the system helps users gradually develop good health behavior habits. The above changes in health behaviors are of great significance for the prevention and control of chronic diseases.

[0073] In summary, the above-mentioned technical solution provides personalized diet and exercise plan recommendations for registered target users by utilizing the simulation and prediction capabilities of the digital twin model, thereby achieving precise, personalized and intelligent health management.

[0074] The embodiment of the present invention proposes a diet and exercise recommendation system based on digital twins, such as Figure 2As shown, the diet and exercise recommendation system based on digital twins includes: A digital twin model building module, used to build a digital twin model using the collected initial physiological data and dynamic physiological data of the user; The data acquisition module is used to collect the initial physiological data of the registered target user after the digital twin model is built, and dynamically adjust and collect the dynamic physiological data of the target user according to the dynamic change state of the dynamic physiological data of the target user, so as to obtain the initial physiological data and dynamic physiological data of the registered target user; The plan recommendation module is used to use the digital twin model to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user.

[0075] The working principle of the above technical solution is: retrieve the user's historical physiological data from the database, including static initial physiological data (such as age, gender, weight, height and basal metabolism) and dynamic physiological data (such as heart rate, blood pressure and exercise calories consumed). The above data is the basis for building a digital twin model. Cleaning the retrieved data is an important data quality control process. By removing invalid data (such as data with too many missing values ​​or data that is obviously inconsistent with common sense), duplicate data and erroneous data, the accuracy and reliability of subsequent model construction are ensured. Use the preprocessed data to build a digital twin model. Train and verify the constructed digital twin model. The training process is to adjust the model parameters through a large amount of data so that the model can more accurately reflect the actual situation; the verification process is to evaluate the performance of the model on unseen data to ensure the generalization ability and stability of the model.

[0076] The effect of the above technical solution is: the digital twin model constructed based on the user's initial physiological data and dynamic physiological data can accurately reflect the physiological characteristics and changing trends of each user, and provide a highly personalized basis for subsequent diet and exercise recommendations. Through data cleaning and model training verification, the accuracy and reliability of the digital twin model are effectively improved, enabling it to more accurately predict the user's physiological response under different diet and exercise plans. The model can receive and process the user's dynamic physiological data in real time, and dynamically adjust the diet and exercise recommendation plan according to changes in the user's physiological state to ensure the scientificity and timeliness of the recommendations. The diet and exercise recommendation method based on the digital twin model can provide users with more accurate and effective health management services, help users better manage their health status, prevent the occurrence of diseases, and improve the quality of life.

[0077] In summary, the above technical solution achieves accurate management and effective intervention of users' health status by building a highly personalized digital twin model and combining dynamic physiological data to update recommended solutions in real time.

[0078] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A diet and exercise recommendation method based on digital twins, characterized in that: The diet and exercise recommendation method based on digital twins includes: Build a digital twin model using the collected initial physiological data and dynamic physiological data of the user; After the digital twin model is built, the initial physiological data of the registered target user is collected, and the dynamic physiological data of the target user is dynamically adjusted and collected according to the dynamic change state of the dynamic physiological data of the target user, so as to obtain the initial physiological data and dynamic physiological data of the registered target user; The digital twin model is used to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user.

2. The method for recommending diet and exercise based on digital twins according to claim 1, characterized in that: The digital twin model is constructed using the collected initial physiological data and dynamic physiological data of the user, including: Retrieving the collected initial physiological data and dynamic physiological data of the user from the database; wherein the initial physiological data includes age, gender, weight, height and basal metabolism; the dynamic physiological data includes heart rate, blood pressure and exercise calories consumed; Cleaning the collected initial physiological data and dynamic physiological data of the user to remove invalid data, duplicate data and erroneous data, and obtaining the pre-processed initial physiological data and dynamic physiological data; A digital twin model is constructed using the preprocessed initial physiological data and dynamic physiological data, and the constructed digital twin model is trained and verified to obtain a trained and verified digital twin model.

3. The method for recommending diet and exercise based on digital twins according to claim 2, characterized in that: The structure of the digital twin model is as follows: An input layer, used to input the initial physiological data and dynamic physiological data of the target user, and perform unified standardization processing on the initial physiological data and dynamic physiological data of the target user; A multi-scale time window layer, including multiple parallel LSTM sub-networks, and each LSTM sub-network corresponds to a time window; wherein the parallel LSTM sub-networks are used to process input sequences of different lengths to control the digital twin model to capture data patterns at different time scales; The adaptive gated fusion layer is used to dynamically adjust the parameter values ​​corresponding to the importance and contribution of each LSTM sub-network according to the current input initial physiological data and dynamic physiological data, and synthesize the output of the dynamically adjusted LSTM sub-network to form a unified hidden layer; A context-aware fusion layer that uses contextual information from the target user to combine with the output of the adaptive gating fusion layer; this dynamically adjusts the model’s response to make the output more personalized and precise; Output layer, used to output diet and running types.

4. The method for recommending diet and exercise based on digital twins according to claim 3, characterized in that: The steps for obtaining the time window are as follows: Retrieve the preset basic decay rate from the database; Real-time monitoring of the frequency of data changes of the collected dynamic physiological data; The dynamic attenuation factor corresponding to each LSTM sub-network is obtained by using the preset basic attenuation rate and the data change frequency of the dynamic physiological data; The dynamic attenuation factor is used to obtain the time length of the time window corresponding to each LSTM sub-network; wherein the time length of the time window corresponding to each LSTM sub-network is obtained by the following formula: Among them, K i represents the time length of the time window corresponding to the i-th LSTM sub-network; T represents the preset observation period length; α(i, β, λ) represents the dynamic attenuation factor corresponding to the i-th LSTM sub-network; N represents the total number of LSTM sub-networks; w i represents the preset initial weight value corresponding to the i-th LSTM sub-network; α0 represents the preset basic attenuation rate; β represents the change frequency factor of dynamic physiological data; i represents the index corresponding to each LSTM sub-network; λ represents the preset periodic intensity value of dynamic physiological data.

5. The method for recommending diet and exercise based on digital twins according to claim 4, characterized in that: The dynamic attenuation factor corresponding to each LSTM sub-network is obtained by using the preset basic attenuation rate and the data change frequency of the dynamic physiological data, including: Retrieve the data change frequency of the collected dynamic physiological data; The change frequency factor is obtained by using the data change frequency of the dynamic physiological data; wherein the change frequency factor is obtained by the following formula: Wherein, β represents the frequency factor of the dynamic physiological data; r represents the adjustment parameter, and the range of the adjustment parameter is [0.1, 0.9]; f 01 、f 02 and f 03 They correspond to the change ratios of the maximum changes in heart rate, blood pressure and exercise calories consumed; v p01 、v p02 and v p03 They correspond to the average rate of change of heart rate, blood pressure and exercise calories consumed per unit time; The dynamic attenuation factor corresponding to each LSTM sub-network is obtained by combining the change frequency factor with the basic attenuation rate; The dynamic attenuation factor corresponding to each LSTM sub-network is obtained by the following formula: Among them, α(i, β, λ) represents the dynamic attenuation factor corresponding to the i-th LSTM sub-network; α0 represents the preset basic attenuation rate; β represents the change frequency factor of the dynamic physiological data; i represents the index corresponding to each LSTM sub-network, and i=1,2,3,…,N; N represents the total number of LSTM sub-networks; λ represents the preset periodic intensity value of the dynamic physiological data.

6. The method for recommending diet and exercise based on digital twin according to claim 1, characterized in that: After the digital twin model is built, the initial physiological data of the registered target user is collected, and the dynamic physiological data of the target user is dynamically adjusted and collected according to the dynamic change state of the dynamic physiological data of the target user, and the initial physiological data and dynamic physiological data of the registered target user are obtained, including: After the digital twin model is built, an initial physiological data input instruction is sent to the registered target user, and the initial physiological data input by the registered target user is received in real time; wherein the initial physiological data includes age, gender, weight, height and basal metabolism; The dynamic physiological data of the registered target user is collected in real time through a wearable device worn by the registered target user; wherein the dynamic physiological data includes heart rate, blood pressure and exercise calories consumed; The change frequency of the dynamic physiological data of the registered target user is extracted to dynamically adjust the collection frequency of the dynamic physiological data.

7. The method for recommending diet and exercise based on digital twins according to claim 6, characterized in that: Extracting the change frequency of the dynamic physiological data of the registered target user to dynamically adjust the collection frequency of the dynamic physiological data includes: Extracting the change frequency of heart rate, blood pressure and exercise calorie consumption contained in the dynamic physiological data of the registered target user; Using the change frequencies of the heart rate, blood pressure and exercise calories consumed, obtaining the change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed; The acquisition frequency of the dynamic physiological data is adjusted using the change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed, wherein the adjusted acquisition frequency of the dynamic physiological data is obtained by the following formula: Among them, f new represents the adjusted acquisition frequency of dynamic physiological data; f y represents the acquisition frequency of dynamic physiological data before adjustment; ξ represents the preset adjustment coefficient, with a value range of 0.26-0.41; X 01 Indicates the frequency parameter corresponding to the heart rate; X 02 Indicates the frequency parameter of blood pressure change; X 03 Indicates the frequency of change corresponding to the calories consumed during exercise.

8. The method for recommending diet and exercise based on digital twins according to claim 7, characterized in that: Using the change frequencies of the heart rate, blood pressure and exercise calories consumed to obtain the change frequency parameters corresponding to the heart rate, blood pressure and exercise calories consumed, includes: The frequency of change of the heart rate included in the dynamic physiological data is used to obtain the frequency of change parameter corresponding to the heart rate; wherein the frequency of change parameter corresponding to the heart rate is obtained by the following formula: Among them, X 01 Indicates the frequency parameter corresponding to the heart rate; X b01 Indicates the standard deviation of heart rate; m 01 Indicates the number of heart rate collections; H i represents the heart rate value corresponding to the i-th acquisition; H i+1 Indicates the heart rate value corresponding to the i+1th acquisition; H f Indicates the maximum allowable fluctuation range of heart rate value; The frequency of change of the blood pressure contained in the dynamic physiological data is used to obtain the frequency parameter of change corresponding to the blood pressure; wherein the frequency parameter of change corresponding to the blood pressure is obtained by the following formula: Among them, X 02 Indicates the frequency parameter of blood pressure change; X x02 represents the standard deviation of diastolic pressure; X y02 represents the standard deviation of systolic blood pressure; m 02 Indicates the number of blood pressure collections; S 01i+1 and S 01i represents the value corresponding to the diastolic pressure collected at the i+1th and ith times; S 02i+1 and S 02i Indicates the corresponding values ​​of systolic blood pressure collected at the i+1th and ith times; The frequency of change of the exercise calorie consumption contained in the dynamic physiological data is used to obtain the frequency parameter of change corresponding to the exercise calorie consumption; wherein the frequency parameter of change corresponding to the exercise calorie consumption is obtained by the following formula: Among them, X 03 Indicates the frequency parameter corresponding to the amount of calories consumed during exercise; X b03 Indicates the standard deviation of calories burned during exercise; m 03 Indicates the number of times the calories consumed during exercise are collected; E i Indicates the exercise calorie consumption value corresponding to the i-th collection; E i+1 Indicates the exercise calorie consumption value corresponding to the i+1th collection; E f Indicates the preset floating reference value for exercise calorie consumption.

9. The method for recommending diet and exercise based on digital twins according to claim 1, characterized in that: The digital twin model is used to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user, including: Using the digital twin model to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, and obtaining a diet and running type that matches the initial physiological data and dynamic physiological data of the registered target user; According to the diet and running type matching the initial physiological data and dynamic physiological data of the registered target user, a diet and exercise plan is retrieved from the database and recommended to the registered target user.

10. A diet and exercise recommendation system based on digital twins, characterized in that: The diet and exercise recommendation system based on digital twins includes: A digital twin model building module, used to build a digital twin model using the collected initial physiological data and dynamic physiological data of the user; The data acquisition module is used to collect the initial physiological data of the registered target user after the digital twin model is built, and dynamically adjust and collect the dynamic physiological data of the target user according to the dynamic change state of the dynamic physiological data of the target user, so as to obtain the initial physiological data and dynamic physiological data of the registered target user; The plan recommendation module is used to use the digital twin model to simulate and predict the initial physiological data and dynamic physiological data of the registered target user, obtain a diet and exercise plan that matches the diet and running type of the registered target user, and recommend it to the registered target user.

Images