Intelligent health state analysis system and method based on artificial intelligence

Abstract

The invention relates to the technical field of health state analysis, in particular to an intelligent health state analysis system and method based on artificial intelligence, and the method comprises the following steps: S1, collecting multi-modal health data; s2, data preprocessing and anomaly detection: preprocessing the collected multi-modal data; s3, multi-modal data fusion and feature selection: carrying out feature fusion on data of different modals through a deep learning model, and carrying out feature extraction on time sequence data by adopting a convolutional neural network (CNN); s4, health state prediction and trend analysis: performing prediction analysis on the health state of the user based on a long and short term memory network (LSTM), and generating a health state trend chart of the user in combination with historical health data and real-time data; and S5, personalized health intervention suggestions are generated: personalized health management suggestions are automatically generated, a more comprehensive health assessment result is provided, the user can better know the own health condition, and personalized management is carried out.

Description

Technical Field

[0001] The present invention relates to the technical field of health status analysis, and in particular to an artificial intelligence-based health status intelligent analysis system and method. Background Art

[0002] With the rapid development of social economy and the continuous improvement of living standards, people are paying more and more attention to their own health. Although traditional health status assessment methods, such as regular physical examinations and doctor consultations, play an important role in health management, they have defects such as untimely data updates and difficulty in real-time monitoring of health changes. In particular, for some potential health problems, traditional methods are often difficult to detect in time, resulting in users failing to take effective preventive measures in the early stages.

[0003] In recent years, with the development of IoT devices, wearable devices and smart medical technologies, the collection of health data has become more convenient. Wearable devices such as smart bracelets and smart watches can monitor users' vital signs and daily behavior data in real time, providing a lot of valuable information for health management. However, many current health analysis systems are limited to analyzing a single type of data and fail to effectively integrate multimodal data (such as vital signs, behavioral data, and environmental data), resulting in a lack of comprehensiveness in health assessment results, making it difficult to provide accurate health status predictions and personalized health intervention recommendations.

[0004] In addition, most existing health prediction systems use traditional linear prediction models, which are unable to handle complex dynamic changes in health status, especially in processing long-term dependencies of time series data. Summary of the invention

[0005] The present invention provides an artificial intelligence-based health status intelligent analysis system and method.

[0006] The health status intelligent analysis method based on artificial intelligence includes the following steps:

[0007] S1: Multimodal health data collection: In the user's daily life, multimodal health data is collected through wearable devices, mobile devices or health monitoring devices. The data includes vital sign data (such as heart rate, blood pressure, body temperature, etc.), behavioral data (such as exercise steps, sleep quality, etc.) and environmental data (such as air quality, temperature and humidity, etc.). Through the joint collection of multiple data sources, the comprehensiveness and real-time nature of the data are ensured;

[0008] S2: Data preprocessing and anomaly detection: Preprocess the collected multimodal data, including data cleaning, standardization and feature extraction, remove redundant information and outliers, and ensure data accuracy;

[0009] S3: Multimodal data fusion and feature selection: The deep learning model is used to fuse the features of data of different modalities, and the convolutional neural network (CNN) is used to extract features from time series data. According to the specific needs of the health analysis task, the feature selection algorithm is used to select the most relevant features and input them into the subsequent analysis module;

[0010] S4: Health status prediction and trend analysis: Based on the long short-term memory network (LSTM), the user's health status is predicted and analyzed, and historical health data and real-time data are combined to generate a user's health status trend chart to predict health problems in the future;

[0011] S5: Generation of personalized health intervention recommendations: Based on predicted health status and trends, combined with medical big data and expert knowledge base, personalized health management recommendations are automatically generated, including diet adjustments, exercise plans, and medical testing recommendations.

[0012] Optionally, the S1 includes:

[0013] S11: Use smart watches or smart bracelets for continuous vital sign monitoring, collecting the user's heart rate, blood pressure and body temperature data in real time;

[0014] S12: Obtaining user behavior data through a mobile device, where the behavior data includes the user's daily step count, daily exercise time, daily sleep duration, and daily calorie consumption;

[0015] S13: Collect external environmental data using an environmental monitoring sensor, the data including PM2.5 concentration in the air, ambient temperature, relative humidity, and ultraviolet radiation intensity;

[0016] S14: All data are transmitted to the cloud data processing center through the Bluetooth communication protocol to ensure that the data is synchronized within 5 seconds after collection, and the security of transmission is ensured by encryption algorithm to avoid data loss or delay.

[0017] Optionally, S2 includes:

[0018] S21: Clean the collected data to remove redundant data caused by sensor failure or the user not wearing the device, to ensure that the input data is complete and correct;

[0019] S22: Use the Z-score standardization method to convert all collected data into a standard format between 0 and 1 to ensure that data from different sources can be processed uniformly;

[0020] S23: Use the 3σ principle based on statistics to identify abnormal points in health data. Abnormal data detected with a heart rate exceeding 180 beats / minute, blood pressure exceeding 200 / 120 mmHg, or body temperature exceeding 40°C will be marked and removed from the data set;

[0021] S24: The detected abnormal data points are corrected using the nearest neighbor interpolation algorithm to ensure that the missing or erroneous data are replaced with reasonable data values ​​consistent with adjacent moments.

[0022] Optionally, the S3 includes:

[0023] S31: Use a three-layer convolutional neural network (CNN) to extract features from the user's heart rate, blood pressure and body temperature time series data. The specific features extracted include the frequency of heart rate fluctuations, the daily variation pattern of blood pressure and the dynamic trend of body temperature.

[0024] S32: Use the Lasso regression algorithm to extract features closely related to the user's health status from multimodal data, remove irrelevant features that have little impact on the analysis results, reduce the system's computational complexity and improve analysis efficiency;

[0025] S33: Use the feature weighted fusion method to fuse vital sign data, behavioral data, and environmental data into a unified health status feature vector, where the weights of heart rate, blood pressure, and air quality are set to 0.4, 0.3, and 0.2, respectively, to ensure that data with high importance receive more attention in the analysis.

[0026] Optionally, the S4 includes:

[0027] S41: Use a long short-term memory network (LSTM) model consisting of 100 neurons to perform time series prediction on the user's heart rate, blood pressure and body temperature data to predict the user's health status trend in the next 24 hours;

[0028] S42: Based on the historical health data of the past 90 days and the real-time collected health data, a health status trend report for the next 7 days is generated. The report content includes the rising or falling trend of the user's heart rate, the change cycle of blood pressure, and the fluctuation pattern of body temperature;

[0029] S43: The system uses a health status trend chart to show the user the changes in heart rate, blood pressure and body temperature in the next 7 days in the form of a line chart. If a risk change exceeding the confidence interval of 75% is detected, the system will send a warning notification to the user;

[0030] S44: The confidence of the prediction results is evaluated through the Bayesian model. When the confidence of the predicted health status is lower than 85%, the system automatically adjusts the parameters of the LSTM model to ensure that the prediction results are sufficiently reliable.

[0031] Optionally, the S5 includes:

[0032] S51: Based on the predicted results of the user's health status, combined with 5 million historical health management data from the medical big data platform and the advice model of 100 health experts, a personalized health management plan including daily diet plan, weekly exercise goals and monthly health check schedule is generated for the user;

[0033] S52: The system dynamically adjusts the intervention strategy based on the personal health goals set by the user, such as losing 5 kg or lowering blood pressure to 120 / 80 mmHg, and provides the user with suggestions for adjusting daily dietary intake and exercise plans in stages over the next 7 days;

[0034] S53: The system automatically combines the user's real-time feedback data, including the execution rate of a certain diet plan or the completion of daily exercise goals, to optimize the health management plan to make it more in line with the user's individual needs;

[0035] S54: Users can receive real-time health management suggestions through mobile applications. If users fail to complete the recommended health tasks as planned, a reminder message will be pushed before the end of the day to urge users to adjust their health management plans.

[0036] Optionally, the three-layer convolutional neural network specifically includes:

[0037] Convolution layer: extracts local features of time series data. The convolution operation multiplies the input data by a "filter" point by point, and then summarizes the results. The convolution of time series data is expressed as: Z = X*W+b, where Z is the result after convolution, Z is the input data (time series such as heart rate and blood pressure), W is the convolution kernel (filter), b is the bias value, and * represents the convolution operation;

[0038] Activation function ReLU: used to introduce nonlinearity: PeLU(Z)=max(0,Z), setting values ​​less than zero to 0 and values ​​greater than zero unchanged;

[0039] Pooling layer: used to reduce dimension and extract important information. Max pooling reduces the amount of data while retaining the most significant features by selecting the maximum value in the pooling window: P = max(X).

[0040] Optionally, the long short-term memory network model controls the information flow through an input gate, a forget gate, and an output gate to ensure that long-term dependencies can be remembered, specifically including:

[0041] Forget gate: decides which unimportant information to delete from memory:

[0042] f t=σ(W f ·x t ), where f t is the output of the forget gate, W f is the weight, x t is the current input data;

[0043] Input gate: decides what new information needs to be stored in memory:

[0044] i t =σ(W i ·x t ), where i t is the output of the input gate, W i is the weight;

[0045] Output gate: controls the output and hidden state update at the current moment:

[0046] o t =σ(W o ·x t );

[0047] Memory state update: The new memory state is controlled by the forget gate and the input gate:

[0048] C t =f t ·C t-1 +i t tanh(W C ·x t ), where C t is the updated memory state, C t-1 It is the memory state of the previous moment;

[0049] Hidden state update: The final hidden state is calculated through the output gate and the current memory state: h t =o t tanh(C t ).

[0050] The health status intelligent analysis system based on artificial intelligence is used to implement the above-mentioned health status intelligent analysis method based on artificial intelligence, and includes the following modules:

[0051] Multimodal health data collection module: used to collect multimodal health data in the user's daily life through wearable devices, mobile devices and health monitoring devices;

[0052] Data preprocessing module: used to preprocess the collected multimodal data, including data cleaning, standardization and feature extraction;

[0053] Multimodal data fusion and feature selection model: used to fuse the features of deep learning models for data of different modalities, and use convolutional neural networks to extract features from time series data;

[0054] Health status prediction and trend analysis module: used to predict and analyze the user's health status based on the long short-term memory network, combine historical health data with real-time collected data, generate a user health status trend chart, and predict health problems that may occur in the future;

[0055] Personalized health intervention recommendation generation module: Based on the health status prediction results, combined with medical big data and expert knowledge base, personalized health management recommendations are automatically generated. The recommendations include diet adjustments, exercise plans and medical testing recommendations to help users optimize their health management methods.

[0056] Beneficial effects of the present invention:

[0057] The present invention, through multimodal data collection and fusion, integrates the user's vital signs data (such as heart rate, blood pressure, body temperature, etc.), behavioral data (such as exercise steps, sleep quality, etc.) and environmental data (such as air quality, temperature and humidity, etc.), ensuring the comprehensiveness and multidimensionality of the data source. Compared with the existing technology, it can effectively integrate data of different modes, use convolutional neural networks to extract features in each mode, and then extract the key features that have the greatest impact on health status through feature selection and weighted fusion. This multimodal data fusion method not only improves the accuracy of health status assessment, but also provides more comprehensive health assessment results, which helps users better understand their own health status and conduct personalized management.

[0058] The present invention uses a long short-term memory network (LSTM) model to perform short-term and long-term time series predictions on the user's health status. The LSTM model can retain the long-term dependencies in the health status data, making the prediction results more accurate and stable. The system combines historical data with real-time data to generate future trend forecasts of the user's health status, especially for dynamic change indicators such as heart rate, blood pressure, and body temperature. It can predict possible future health problems in advance and send reminders to users in a timely manner through trend charts and early warning notifications. Compared with existing static health analysis methods, it can dynamically capture health change trends, significantly improve the initiative and predictability of health management, and help users prevent potential health risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0062] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0063] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0064] like Figure 1 As shown, the health status intelligent analysis method based on artificial intelligence includes the following steps:

[0065] S1: Multimodal health data collection: In the daily life of users, multimodal health data is collected through wearable devices, mobile devices or health monitoring devices. The data includes vital signs data (such as heart rate, blood pressure, body temperature, etc.), behavioral data (such as exercise steps, sleep quality, etc.) and environmental data (such as air quality, temperature and humidity, etc.). Through the joint collection of multiple data sources, the comprehensiveness and real-time nature of the data are ensured;

[0066] S2: Data preprocessing and anomaly detection: Preprocess the collected multimodal data, including data cleaning, standardization and feature extraction, remove redundant information and outliers, and ensure data accuracy;

[0067] S3: Multimodal data fusion and feature selection: The deep learning model is used to fuse the features of data of different modalities, and the convolutional neural network (CNN) is used to extract features from time series data. According to the specific needs of the health analysis task, the feature selection algorithm is used to select the most relevant features and input them into the subsequent analysis module;

[0068] S4: Health status prediction and trend analysis: Based on the long short-term memory network (LSTM), the user's health status is predicted and analyzed, and historical health data and real-time data are combined to generate a user's health status trend chart to predict health problems in the future;

[0069] S5: Generation of personalized health intervention recommendations: Based on predicted health status and trends, combined with medical big data and expert knowledge base, personalized health management recommendations are automatically generated. Recommendations include diet adjustments, exercise plans, and medical testing recommendations to help users adjust their lifestyles in a timely manner and prevent health problems.

[0070] S1 includes:

[0071] S11: Use smart watches or smart bracelets for continuous vital sign monitoring, collecting the user's heart rate, blood pressure and body temperature data in real time;

[0072] S12: Obtaining user behavior data through a mobile device, where the behavior data includes the user's daily step count, daily exercise time, daily sleep duration, and daily calorie consumption;

[0073] S13: Use environmental monitoring sensors to collect external environmental data, including PM2.5 concentration in the air, ambient temperature, relative humidity, and ultraviolet radiation intensity;

[0074] S14: All data are transmitted to the cloud data processing center through the Bluetooth communication protocol to ensure that the data is synchronized within 5 seconds after collection, and the security of transmission is ensured by encryption algorithm to avoid data loss or delay.

[0075] S2 includes:

[0076] S21: Clean the collected data to remove redundant data caused by sensor failure or the user not wearing the device, to ensure that the input data is complete and correct;

[0077] S22: Use the Z-score standardization method to convert all collected data into a standard format between 0 and 1 to ensure that data from different sources can be processed uniformly;

[0078] S23: Use the 3σ principle based on statistics to identify abnormal points in health data. Abnormal data detected with a heart rate exceeding 180 beats / minute, blood pressure exceeding 200 / 120 mmHg, or body temperature exceeding 40°C will be marked and removed from the data set;

[0079] S24: The detected abnormal data points are corrected using the nearest neighbor interpolation algorithm to ensure that the missing or erroneous data are replaced with reasonable data values ​​consistent with adjacent moments.

[0080] S3 includes:

[0081] S31: Use a three-layer convolutional neural network (CNN) to extract features from the user's heart rate, blood pressure and body temperature time series data. The specific features extracted include the frequency of heart rate fluctuations, the daily variation pattern of blood pressure and the dynamic trend of body temperature.

[0082] S32: Use the Lasso regression algorithm to extract features closely related to the user's health status from multimodal data, remove irrelevant features that have little impact on the analysis results, reduce the system's computational complexity and improve analysis efficiency;

[0083] S33: Use the feature weighted fusion method to fuse vital sign data, behavioral data, and environmental data into a unified health status feature vector, where the weights of heart rate, blood pressure, and air quality are set to 0.4, 0.3, and 0.2, respectively, to ensure that data with high importance receive more attention in the analysis.

[0084] S4 includes:

[0085] S41: Use a long short-term memory network (LSTM) model consisting of 100 neurons to perform time series prediction on the user's heart rate, blood pressure and body temperature data to predict the user's health status trend in the next 24 hours;

[0086] S42: Based on the historical health data of the past 90 days and the real-time collected health data, a health status trend report for the next 7 days is generated. The report content includes the rising or falling trend of the user's heart rate, the change cycle of blood pressure, and the fluctuation pattern of body temperature;

[0087] S43: The system uses a health status trend chart to show the user the changes in heart rate, blood pressure and body temperature in the next 7 days in the form of a line chart. If a risk change exceeding the confidence interval of 75% is detected, the system will send a warning notification to the user;

[0088] S44: The confidence of the prediction results is evaluated through the Bayesian model. When the confidence of the predicted health status is lower than 85%, the system automatically adjusts the parameters of the LSTM model to ensure that the prediction results are sufficiently reliable.

[0089] S5 includes:

[0090] S51: Based on the predicted results of the user's health status, combined with 5 million historical health management data from the medical big data platform and the advice model of 100 health experts, a personalized health management plan including daily diet plan, weekly exercise goals and monthly health check schedule is generated for the user;

[0091] S52: The system dynamically adjusts the intervention strategy based on the personal health goals set by the user, such as losing 5 kg or lowering blood pressure to 120 / 80 mmHg, and provides the user with suggestions for adjusting daily dietary intake and exercise plans in stages over the next 7 days;

[0092] S53: The system automatically combines the user's real-time feedback data, including the execution rate of a certain diet plan or the completion of daily exercise goals, to optimize the health management plan to make it more in line with the user's individual needs;

[0093] S54: Users can receive real-time health management suggestions through mobile applications. If users fail to complete the recommended health tasks as planned, a reminder message will be pushed before the end of the day to urge users to adjust their health management plans.

[0094] The three-layer convolutional neural network specifically includes:

[0095] Convolution layer: extracts local features of time series data. The convolution operation multiplies the input data by a "filter" point by point, and then summarizes the results. The convolution of time series data is expressed as: Z = X*W+b, where Z is the result after convolution, X is the input data (time series such as heart rate and blood pressure), W is the convolution kernel (filter), b is the bias value, and * represents the convolution operation;

[0096] Activation function ReLU: used to introduce nonlinearity: ReLU(Z)=max(0,Z), sets values ​​less than zero to 0 and values ​​greater than zero to remain unchanged;

[0097] Pooling layer: used to reduce dimension and extract important information. Max pooling reduces the amount of data while retaining the most significant features by selecting the maximum value in the pooling window: P = max(X).

[0098] The LSTM network model controls the information flow through the input gate, forget gate, and output gate to ensure that long-term dependencies can be remembered, including:

[0099] Forget gate: decides which unimportant information to delete from memory:

[0100] f t =σ(W f ·x t ), where f t is the output of the forget gate, W f is the weight, x t is the current input data;

[0101] Input gate: decides what new information needs to be stored in memory:

[0102] i t =σ(W i ·x i ), where i t is the output of the input gate, W i is the weight;

[0103] Output gate: controls the output and hidden state update at the current moment:

[0104] o t =σ(W o ·x t );

[0105] Memory state update: The new memory state is controlled by the forget gate and the input gate:

[0106] C t =f t ·C t-1 +i t tanh(W C ·x t ), where C t is the updated memory state, X t-1 It is the memory state of the previous moment;

[0107] Hidden state update: The final hidden state is calculated through the output gate and the current memory state: h t =o t tanh(C t ).

[0108] The health status intelligent analysis system based on artificial intelligence is used to implement the above-mentioned health status intelligent analysis method based on artificial intelligence, and includes the following modules:

[0109] Multimodal health data collection module: used to collect multimodal health data in the user's daily life through wearable devices, mobile devices and health monitoring devices;

[0110] Data preprocessing module: used to preprocess the collected multimodal data, including data cleaning, standardization and feature extraction;

[0111] Multimodal data fusion and feature selection model: used to fuse the features of deep learning models for data of different modalities, and use convolutional neural networks to extract features from time series data;

[0112] Health status prediction and trend analysis module: used to predict and analyze the user's health status based on the long short-term memory network, combine historical health data with real-time collected data, generate a user health status trend chart, and predict health problems that may occur in the future;

[0113] Personalized health intervention recommendation generation module: Based on the health status prediction results, combined with medical big data and expert knowledge base, personalized health management recommendations are automatically generated. The recommendations include diet adjustments, exercise plans and medical testing recommendations to help users optimize their health management methods.

[0114] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0115] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An intelligent analysis method of health status based on artificial intelligence, characterized in that: The following steps are involved: S1: Multimodal health data collection: In the user's daily life, multimodal health data is collected through wearable devices, mobile devices or health monitoring devices. The data includes vital signs data, behavioral data and environmental data. The comprehensiveness and real-time nature of the data are ensured through the joint collection of multiple data sources; S2: Data preprocessing and anomaly detection: Preprocess the collected multimodal data, including data cleaning, standardization and feature extraction, remove redundant information and outliers, and ensure data accuracy; S3: Multimodal data fusion and feature selection: The deep learning model is used to fuse the features of data of different modalities, and the convolutional neural network is used to extract features from time series data. According to the specific needs of the health analysis task, the feature selection algorithm is used to select the most relevant features and input them into the subsequent analysis module; S4: Health status prediction and trend analysis: Based on the long short-term memory network, the user's health status is predicted and analyzed. The historical health data and real-time data are combined to generate the user's health status trend chart to predict health problems in the future. S5: Generation of personalized health intervention recommendations: Based on predicted health status and trends, combined with medical big data and expert knowledge base, personalized health management recommendations are automatically generated, including diet adjustments, exercise plans, and medical testing recommendations.

2. The method for intelligent analysis of health status based on artificial intelligence according to claim 1, characterized in that: The S1 includes: S11: Use smart watches or smart bracelets for continuous vital sign monitoring, collecting the user's heart rate, blood pressure and body temperature data in real time; S12: Obtaining user behavior data through a mobile device, where the behavior data includes the user's daily step count, daily exercise time, daily sleep duration, and daily calorie consumption; S13: Collect external environmental data using an environmental monitoring sensor, the data including PM2.5 concentration in the air, ambient temperature, relative humidity, and ultraviolet radiation intensity; S14: All data are transmitted to the cloud data processing center through the Bluetooth communication protocol to ensure that the data is synchronized within 5 seconds after collection, and the security of transmission is ensured by encryption algorithm to avoid data loss or delay.

3. The method for intelligent analysis of health status based on artificial intelligence according to claim 1, characterized in that: The S2 includes: S21: Clean the collected data to remove redundant data caused by sensor failure or the user not wearing the device, to ensure that the input data is complete and correct; S22: Use the Z-score standardization method to convert all collected data into a standard format between 0 and 1 to ensure that data from different sources can be processed uniformly; S23: Use the 3σ principle based on statistics to identify abnormal points in health data. Abnormal data detected with a heart rate exceeding 180 beats / minute, blood pressure exceeding 200 / 120 mmHg, or body temperature exceeding 40°C will be marked and removed from the data set; S24: The detected abnormal data points are corrected using the nearest neighbor interpolation algorithm to ensure that the missing or erroneous data are replaced with reasonable data values ​​consistent with adjacent moments.

4. The method for intelligent analysis of health status based on artificial intelligence according to claim 1, characterized in that: The S3 includes: S31: Use a three-layer convolutional neural network to extract features from the user's heart rate, blood pressure and body temperature time series data. The specific features extracted include the frequency of heart rate fluctuations, the daily variation pattern of blood pressure and the dynamic trend of body temperature. S32: Use the Lasso regression algorithm to extract features closely related to the user's health status from multimodal data, remove irrelevant features that have little impact on the analysis results, reduce the system's computational complexity and improve analysis efficiency; S33: Use the feature weighted fusion method to fuse vital sign data, behavioral data, and environmental data into a unified health status feature vector, where the weights of heart rate, blood pressure, and air quality are set to 0.4, 0.3, and 0.2, respectively, to ensure that data with high importance receive more attention in the analysis.

5. The method for intelligent analysis of health status based on artificial intelligence according to claim 1, characterized in that: The S4 includes: S41: Use a long short-term memory network model consisting of 100 neurons to perform time series prediction on the user's heart rate, blood pressure and body temperature data to predict the user's health status trend in the next 24 hours; S42: Based on the historical health data of the past 90 days and the real-time collected health data, a health status trend report for the next 7 days is generated. The report content includes the rising or falling trend of the user's heart rate, the change cycle of blood pressure, and the fluctuation pattern of body temperature; S43: The system uses a health status trend chart to show the user the changes in heart rate, blood pressure and body temperature in the next 7 days in the form of a line chart. If a risk change exceeding the confidence interval of 75% is detected, the system will send a warning notification to the user; S44: The confidence of the prediction results is evaluated through the Bayesian model. When the confidence of the predicted health status is lower than 85%, the system automatically adjusts the parameters of the LSTM model to ensure that the prediction results are sufficiently reliable.

6. The method for intelligent analysis of health status based on artificial intelligence according to claim 1, characterized in that: The S5 includes: S51: Based on the predicted results of the user's health status, combined with 5 million historical health management data from the medical big data platform and the advice model of 100 health experts, a personalized health management plan including daily diet plan, weekly exercise goals and monthly health check schedule is generated for the user; S52: The system dynamically adjusts the intervention strategy based on the personal health goals set by the user, such as losing 5 kg or lowering blood pressure to 120 / 80 mmHg, and provides the user with suggestions for adjusting daily dietary intake and exercise plans in stages over the next 7 days; S53: The system automatically combines the user's real-time feedback data, including the execution rate of a certain diet plan or the completion of daily exercise goals, to optimize the health management plan to make it more in line with the user's individual needs; S54: Users can receive real-time health management suggestions through mobile applications. If users fail to complete the recommended health tasks as planned, a reminder message will be pushed before the end of the day to urge users to adjust their health management plans.

7. The method for intelligent analysis of health status based on artificial intelligence according to claim 4, characterized in that: The three-layer convolutional neural network specifically includes: Convolution layer: extracts local features of time series data. The convolution operation multiplies the input data by a "filter" point by point, and then summarizes the results. The convolution of time series data is expressed as: Z = X*W+b, where Z is the result after convolution, X is the input data (time series such as heart rate and blood pressure), W is the convolution kernel (filter), b is the bias value, and * represents the convolution operation; Activation function ReLU: used to introduce nonlinearity: ReLU(Z) = max(0, Z), setting values ​​less than zero to 0 and values ​​greater than zero unchanged; Pooling layer: used to reduce dimension and extract important information. Max pooling reduces the amount of data while retaining the most significant features by selecting the maximum value in the pooling window: P = max(X).

8. The method for intelligent analysis of health status based on artificial intelligence according to claim 5, characterized in that: The LSTM network model controls the information flow through the input gate, forget gate, and output gate to ensure that long-term dependencies can be remembered, including: Forget gate: decides which unimportant information to delete from memory: f t =σ(W f ·x t ), where f t is the output of the forget gate, W f is the weight, x t is the current input data; Input gate: decides what new information needs to be stored in memory: i t =σ(W i ·x t ), where i t is the output of the input gate, W i is the weight; Output gate: controls the output and hidden state update at the current moment: oh t =σ(W0·x t ); Memory state update: The new memory state is controlled by the forget gate and the input gate: C t =f t ·C t-1 +i t tanh(W c ·x t ), where C t is the updated memory state, C t-1 It is the memory state of the previous moment; Hidden state update: The final hidden state is calculated through the output gate and the current memory state: h t =o t tanh(C t ).

9. An artificial intelligence-based health status intelligent analysis system, used to implement the artificial intelligence-based health status intelligent analysis method according to any one of claims 1 to 8, characterized in that: Includes the following modules: Multimodal health data collection module: used to collect multimodal health data in the user's daily life through wearable devices, mobile devices and health monitoring devices; Data preprocessing module: used to preprocess the collected multimodal data, including data cleaning, standardization and feature extraction; Multimodal data fusion and feature selection model: used to fuse the features of deep learning models for data of different modalities, and use convolutional neural networks to extract features from time series data; Health status prediction and trend analysis module: used to predict and analyze the user's health status based on the long short-term memory network, combine historical health data with real-time collected data, generate a user health status trend chart, and predict health problems that may occur in the future; Personalized health intervention recommendation generation module: Based on the health status prediction results, combined with medical big data and expert knowledge base, personalized health management recommendations are automatically generated. The recommendations include diet adjustments, exercise plans and medical testing recommendations to help users optimize their health management methods.

Images