Health monitoring time prediction system based on big data analysis

Abstract

The present invention relates to the field of time prediction technology, and in particular, to a health monitoring time prediction system based on big data analysis. It includes an acquisition analysis unit, a data model unit, an evaluation display unit, and an intervention suggestion unit. The data model unit of the present invention establishes a time prediction model based on analyzed individual health status data, extracted individual health data, and extracted historical individual health data. When receiving data of insufficient data acquisition frequency command data, the data model unit extracts individual living habits, genetic factors, and environmental impact data from the analysis and judgment module to update the time prediction model. The time prediction model can predict the health problems and disease risks faced by individuals in the future, and help individuals take preventive and intervention measures as soon as possible. At the same time, by updating the data in the most recent stage, the timeliness and accuracy of the time prediction model can be maintained, and the current health status and dynamic changes of individuals can be better reflected, and the accuracy and reliability of the prediction can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of time prediction, and in particular to a health monitoring time prediction system based on big data analysis. Background Art

[0002] The health monitoring time prediction based on big data analysis can predict the individual's health status and possible disease risks in the future by collecting a large amount of health data, such as heart rate, body temperature, and exercise status. It can help medical institutions or individuals to discover health problems in advance and take timely intervention measures, thereby improving the efficiency and effectiveness of health management. Since the acquisition of a large amount of health data will lead to data missing, when the acquired health data has data missing, it means that the data has not been acquired or recorded, making the data incomplete and unable to accurately predict the data, resulting in insufficient data collection frequency and reduced data integrity. If the acquired data has insufficient data collection frequency, the low data collection frequency will lead to untimely updates of the prediction model and inability to capture the changing trends and fluctuations of individual health, making it impossible for the prediction model to accurately predict future health status, resulting in errors in the data predicted by the prediction model, reducing the accuracy and reliability of the prediction model. In order to avoid errors in the data predicted by the prediction model due to insufficient data collection frequency, we provide a health monitoring time prediction system based on big data analysis. Summary of the invention

[0003] The purpose of the present invention is to provide a health monitoring time prediction system based on big data analysis to solve the problems raised in the above background technology.

[0004] To achieve the above-mentioned purpose, the present invention provides a health monitoring time prediction system based on big data analysis, including an acquisition and analysis unit, a data model unit, an evaluation and display unit, and an intervention suggestion unit;

[0005] The acquisition and analysis unit acquires individual health data and historical individual health data, performs data quality analysis based on the acquired individual health data and determines whether the data collection frequency is insufficient, and then performs individual health status analysis based on the acquired individual health data;

[0006] The data model unit is used to receive the data in the acquisition analysis unit, the data model unit acquires individual health data and historical individual health data from the analysis and judgment module and establishes a time prediction model and updates the time prediction model;

[0007] The evaluation and display unit is used to receive the time prediction model established in the data model unit and obtain the individual health status data analyzed in the analysis unit and evaluate the individual health index, and then display the evaluated individual health index data to the outside world;

[0008] The intervention suggestion unit is used to receive the data in the data model unit and obtain the historical individual health data from the analysis and judgment module, and make intervention suggestions for lifestyle according to the data in the data model unit and the obtained historical individual health data.

[0009] As a further improvement of the technical solution, the acquisition and analysis unit includes an analysis and judgment module, a data fusion module and a situation analysis module;

[0010] The analysis and judgment module uses the monitoring device to monitor individual health data in real time and obtain individual health data, and obtains historical individual health data at the same time, stores the obtained individual health data and the obtained historical individual health data in a database to establish a complete storage system, and then analyzes the data quality based on the obtained individual health data, and uses the analyzed data quality and the set standard data quality to determine whether the data collection frequency is insufficient;

[0011] Historical individual health data includes environmental data, lifestyle habits, genetic factors, social data, and medical records;

[0012] The data fusion module is used to receive the data collection frequency insufficient command data in the analysis and judgment module. The data fusion module extracts individual health data and historical individual health data from the database in the analysis and judgment module, and fuses the individual health data and the historical individual health data into multi-source data.

[0013] As a further improvement of the technical solution, the specific judgment situations in the analysis and judgment module include:

[0014] Case ①: When the quality of the analyzed data is less than the set standard data quality, it means that there is data missing in the acquired individual health data, resulting in insufficient data collection frequency, and the insufficient data collection frequency command data is transmitted to the data fusion module;

[0015] Case ②: When the quality of the analyzed data is greater than the set standard data quality, it indicates that the acquired individual health data has data duplication, and the acquired individual health data is deduplicated, and the deduplicated individual health data is transferred to the status analysis module;

[0016] Case ③: When the quality of the analyzed data is consistent with the set standard data quality, it means that there is no missing data or data duplication in the acquired individual health data, and the command data of no missing data or data duplication is transmitted to the status analysis module.

[0017] As a further improvement of the present technical solution, the situation analysis module is used to receive command data without data missing and data duplication in the analysis and judgment module or deduplicated individual health data or multi-source data fused in the data fusion module. When command data without data missing and data duplication is received, the situation analysis module extracts individual health data from the database in the analysis and judgment module, and uses K-means in big data analysis to analyze the individual health status of the extracted individual health data or the fused multi-source data or deduplicated individual health data.

[0018] As a further improvement of the technical solution, the implementation steps of using K-means in the status analysis module to analyze the extracted individual health data or the fused multi-source data or the deduplicated individual health data to analyze the individual health status are as follows:

[0019] Step 1: First collect the extracted individual health data GT jk or fused multi-source data DY sj Or deduplicated individual health data QC jk , and then from the extracted individual health data GT jk or fused multi-source data DY sj Or deduplicated individual health data QC jk Get feature T z ;

[0020] Feature T z Including physiological index characteristic data SL zb , motion characteristic data YD tz , Sleep characteristic data SM tz , Eating habits characteristic data YS xg and weight characteristic data TZ tz ;

[0021] Step 2: According to the physiological index characteristic data SL zb , motion characteristic data YD tz , Sleep characteristic data SM tz , Eating habits characteristic data YS xg and weight characteristic data TZ tz Find out the features T that are more correlated with individual health status z ;

[0022] Step ③: According to the characteristics T that are more correlated with individual health status z Determine the K value and select the K value as the initial cluster center JL zx , and then the characteristics T with higher correlation with individual health status are obtained z To each cluster center JL zx The distance between the individual and the health status of the individual is higher. zAssigned to the closest cluster center JL zx , update cluster center JL zx until the convergence condition is reached and the cluster JL is obtained c ;

[0023] Step ④: According to cluster JL c The features combined with domain knowledge zs and experience y Each cluster c is explained, the individual health status represented by each cluster c is explained, and the analyzed individual health status data FX is obtained. jk .

[0024] As a further improvement of the present technical solution, the data model unit is used to receive the individual health status data analyzed in the status analysis module and the data collection frequency insufficient command data in the analysis and judgment module, the data model unit extracts the individual health data and the historical individual health data from the database in the analysis and judgment module, the data model unit uses time series analysis to establish a time prediction model according to the analyzed individual health status data, the extracted individual health data and the extracted historical individual health data, the prediction model is used to collect individual living habits, genetic factors, and environmental impact data, when the data collection frequency insufficient command data is received, the data model unit extracts the individual living habits, genetic factors, and environmental impact data of the most recent stage from the database in the analysis and judgment module to update the time prediction model;

[0025] The implementation steps of building a time prediction model using time series analysis are as follows:

[0026] Step 1: First collect and analyze the individual health status data FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk , and draw the analyzed individual health status data FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk Timing diagram SX over time t ;

[0027] Step 2: Draw the timing diagram SX t FX, the individual health status data for detection and analysis jk , extracted individual health data GT jk and extracted historical individual health data LS jk Is there any autocorrelation? Observe the individual health data of the analysis at a certain point in time. jk , extracted individual health data GT jk and extracted historical individual health data LSjk Whether the value is related to the previous time point;

[0028] Step 3: Observe the timing diagram SX t Does the data show a clear upward or downward trend and draw an observed trend? qs , and then according to the timing diagram SX t Perform seasonal decomposition analysis and obtain seasonal characteristics JJX tz , according to the observed trend GC qs and seasonal characteristics JJX tz FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk Fitting and modeling are performed to obtain the established time prediction model SJ mx ;

[0029] The implementation steps for extracting the most recent individual lifestyle habits, genetic factors, and environmental impact data to update the time prediction model are as follows:

[0030] Step ①: First, receive the data acquisition frequency insufficient command data PL bz Time prediction model SJ mx From the database SJ k Extract the individual lifestyle habits of the most recent stage from SH xg , Genetic factors ys 、Environmental impact yx data;

[0031] Step 2: SH your recent individual lifestyle habits xg , Genetic factors ys 、Environmental impact yx The data is based on the 70% training set XL j and 30% validation set YZ j Divide and then transform the time prediction model SJ mx Applied to training dataset XL sjj , and then the time prediction model SJ mx and training individual living habits SH xg , Genetic factors ys 、Environmental impact yx Match the features in the data and generate the prediction result YC jg ;

[0032] Step 3: Use the time prediction model SJ mx Generated prediction result YC jg With the training dataset XL sjj The actual observed value SJ ingcz Compare and compare the time prediction model SJ mx Output and actual observed value SJ gcz , and observe the time prediction model SJ mx How well it fits the training data;

[0033] Step ④: SJ based on time prediction model mx Output and actual observed value SJ gcz The difference between the adjusted time prediction model SJ mx The parameters of the SJ prediction model are adjusted in time mx After the parameters are set, reapply the time prediction model SJ mx Based on the training data, a new round of time prediction model SJ is performed mx The output is compared with the actual observation value. This process is usually an iterative process until the time prediction model SJ mx Until the performance reaches a satisfactory level, it shows that the time prediction model SJ mx Update completed;

[0034] As a further improvement of the technical solution, the evaluation and display unit includes an indicator evaluation module and a data display module;

[0035] The index evaluation module is used to receive the time prediction model established in the data model unit and the individual health status data analyzed in the status analysis module, and evaluate the individual health index according to the established time prediction model and the analyzed individual health status data to obtain the evaluated individual health index data, and the evaluated individual health index data is used to evaluate the body mass index and cardiovascular health index data;

[0036] The data display module is used to receive the individual health index data evaluated in the index evaluation module and display it to the outside world.

[0037] As a further improvement of the present technical solution, the data model unit receives the individual health indicator data evaluated in the indicator evaluation module, and the updated time prediction model predicts the disease risk of the evaluated individual health indicator data and the extracted historical individual health data, predicts the individual's future risk of cardiovascular disease, diabetes, and cancer, and obtains predicted disease risk data.

[0038] As a further improvement of the present technical solution, the intervention suggestion unit is used to receive the disease risk data predicted in the data model unit. The intervention suggestion unit extracts historical individual health data from the database in the analysis and judgment module, and makes lifestyle intervention suggestions based on the predicted disease risk data and the environmental data, living habits, genetic factors and medical record data in the extracted historical individual health data.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. In the health monitoring time prediction system of the big data analysis, the data model unit receives the individual health status data analyzed in the acquisition analysis unit and the insufficient data collection frequency command data in the analysis and judgment module. The data model unit establishes a time prediction model based on the analyzed individual health status data, the extracted individual health data and the extracted historical individual health data. When the insufficient data collection frequency command data is received, the data model unit extracts the recent individual living habits, genetic factors, and environmental impact data from the database in the analysis and judgment module to update the time prediction model. The time prediction model can predict the health problems and disease risks that individuals will face in the future, and help individuals take preventive and intervention measures as early as possible. At the same time, by updating the recent data, the timeliness and accuracy of the time prediction model can be maintained, better reflecting the current health status and dynamic changes of the individual, and improving the accuracy and reliability of the prediction.

[0041] 2. In the health monitoring time prediction system of big data analysis, the indicator evaluation module evaluates individual health indicators based on the established time prediction model and the analyzed individual health status data. The time prediction model and individual health status data can be used to quantitatively evaluate the individual's health status. Through the analysis and comparison of various health indicators, a more comprehensive understanding of the individual's overall health level can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is an overall block diagram of the present invention;

[0043] Figure 2 The block diagram of the acquisition and analysis unit of the present invention;

[0044] Figure 3 A block diagram of a unit for evaluating the present invention is shown.

[0045] The meaning of each number in the figure is:

[0046] 1. Acquisition and analysis unit; 11. Analysis and judgment module; 12. Data fusion module; 13. Situation analysis module;

[0047] 2. Data model unit; 3. Evaluation and display unit; 31. Index evaluation module; 32. Data display module;

[0048] 4. Intervention suggestion unit. DETAILED DESCRIPTION

[0049] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Example 1

[0051] The present invention provides a health monitoring time prediction system based on big data analysis, please refer to Figure 1-Figure 3 , including acquisition analysis unit 1, data model unit 2, evaluation and presentation unit 3, and intervention suggestion unit 4;

[0052] The acquisition and analysis unit 1 acquires individual health data and historical individual health data, analyzes the data quality based on the acquired individual health data and determines whether the data collection frequency is insufficient, and then analyzes the individual health status based on the acquired individual health data; the data model unit 2 is used to receive the data in the acquisition and analysis unit 1, and the data model unit 2 acquires the individual health data and historical individual health data from the analysis and judgment module 11 and establishes a time prediction model and updates the time prediction model; the evaluation and display unit 3 is used to receive the time prediction model established in the data model unit 2 and the individual health status data analyzed in the acquisition and analysis unit 1 and evaluate the individual health indicators, and then display the evaluated individual health indicator data to the outside world; the intervention suggestion unit 4 is used to receive the data in the data model unit 2 and acquire the historical individual health data from the analysis and judgment module 11, and make lifestyle intervention suggestions based on the data in the data model unit 2 and the acquired historical individual health data.

[0053] The following is a refinement of the above units, see Figure 1-Figure 3 ;

[0054] The acquisition and analysis unit 1 includes an analysis and judgment module 11, a data fusion module 12 and a situation analysis module 13;

[0055] The analysis and judgment module 11 uses the monitoring equipment to monitor individual health data in real time and obtain individual health data, and obtains historical individual health data at the same time, and stores the obtained individual health data and the obtained historical individual health data in the database to establish a complete storage system, and then analyzes the data quality based on the obtained individual health data, and uses the analyzed data quality and the set standard data quality to judge whether the data collection frequency is insufficient. Through data quality analysis, deviations or abnormalities in the data can be found, and it can be judged whether the data collection frequency is sufficient to cover various situations, and corresponding measures can be taken to avoid inaccurate or incomplete analysis results due to insufficient data collection frequency;

[0056] Historical individual health data includes environmental data, lifestyle habits, genetic factors, social data, and medical records;

[0057] The specific judgment situations in the analysis and judgment module 11 include:

[0058] Case ①: When the quality of the analyzed data is less than the set standard data quality, it indicates that the acquired individual health data has data missing, resulting in insufficient data collection frequency, and the insufficient data collection frequency command data is transmitted to the data fusion module 12;

[0059] Case ②: When the quality of the analyzed data is greater than the set standard data quality, it indicates that the acquired individual health data has data duplication, and the acquired individual health data is deduplicated, and the deduplicated individual health data is transferred to the status analysis module 13;

[0060] Case ③: When the quality of the analyzed data is consistent with the set standard data quality, it means that there is no missing data or data duplication in the acquired individual health data, and the command data of no missing data or data duplication is transmitted to the condition analysis module 13;

[0061] The data fusion module 12 is used to receive the data of insufficient data collection frequency command in the analysis and judgment module 11. The data fusion module 12 extracts individual health data and historical individual health data from the database in the analysis and judgment module 11, and fuses the individual health data and the historical individual health data into multi-source data, eliminates redundant information, retains core data, and obtains fused multi-source data. It can combine multiple data sources, conduct comprehensive analysis, fill in missing data, and improve data integrity;

[0062] The condition analysis module 13 is used to receive the command data without data missing and data duplication or the individual health data without duplication or the multi-source data fused in the data fusion module 12 in the analysis and judgment module 11. When the command data without data missing and data duplication is received, the condition analysis module 13 extracts the individual health data from the database in the analysis and judgment module 11, and uses K-means in big data analysis to analyze the individual health status of the extracted individual health data or the fused multi-source data or the individual health data without duplication. By analyzing the individual health data, the individual health status can be comprehensively evaluated, including physiological indicators such as weight, blood pressure, blood sugar, cholesterol, and living habits, eating habits, exercise conditions and other factors, to help the individual understand his or her own health status;

[0063] Use K-means to analyze the extracted individual health data or fused multi-source data or deduplicated individual health data to implement the steps of individual health status:

[0064] Step 1: First collect the extracted individual health data GTjk or fused multi-source data DY sj Or deduplicated individual health data QC jk , and then from the extracted individual health data GT jk or fused multi-source data DY sj Or deduplicated individual health data QC jk Get feature T z ;

[0065] Feature T z Including physiological index characteristic data SL zb , motion characteristic data YD tz , Sleep characteristic data SM tz , Eating habits characteristic data YS xg and weight characteristic data TZ tz ;

[0066] Step 2: According to the physiological index characteristic data SL zb , motion characteristic data YD tz , Sleep characteristic data SM tz , Eating habits characteristic data YS xg and weight characteristic data TZ tz Find out the features T that are more correlated with individual health status z ;

[0067] Step ③: According to the characteristics T that are more correlated with individual health status z Determine the K value and select the K value as the initial cluster center JL zx , and then the characteristics T with higher correlation with individual health status are obtained z To each cluster center JL zx The distance between the individual and the health status of the individual is higher. z Assigned to the closest cluster center JL zx , update cluster center JL zx until the convergence condition is reached and the cluster JL is obtained c ;

[0068] Step ④: According to cluster JL c The features combined with domain knowledge zs and experience y Each cluster c is explained, the individual health status represented by each cluster c is explained, and the analyzed individual health status data FX is obtained. jk ;

[0069] The data model unit 2 is used to receive the individual health status data analyzed in the status analysis module 13 and the data collection frequency insufficient command data in the analysis and judgment module 11. The data model unit 2 extracts the individual health data and the historical individual health data from the database in the analysis and judgment module 11. The data model unit 2 uses time series analysis to establish a time prediction model based on the analyzed individual health status data, the extracted individual health data and the extracted historical individual health data. The prediction model is used to collect individual living habits, genetic factors, and environmental impact data. When the data collection frequency insufficient command data is received, the data model unit 2 extracts the most recent individual living habits, genetic factors, and environmental impact data from the database in the analysis and judgment module 11 to update the time prediction model. The time prediction model can predict the health problems and disease risks that an individual may face in the future, and help the individual take preventive and intervention measures as early as possible. At the same time, by updating the most recent data, the timeliness and accuracy of the time prediction model can be maintained, better reflecting the individual's current health status and dynamic changes, and improving the accuracy and reliability of the prediction.

[0070] The implementation steps of building a time prediction model using time series analysis are as follows:

[0071] Step 1: First collect and analyze the individual health status data FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk , and draw the analyzed individual health status data FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk Timing diagram SX over time t ;

[0072] Step 2: Draw the timing diagram SX t FX, the individual health status data for detection and analysis jk , extracted individual health data GT jk and extracted historical individual health data LS jk Is there any autocorrelation? Observe the individual health data of the analysis at a certain point in time. jk , extracted individual health data GT jk and extracted historical individual health data LS jk Whether the value is related to the previous time point;

[0073] Step 3: Observe the timing diagram SX t Does the data show a clear upward or downward trend and draw an observed trend? qs , and then according to the timing diagram SX tPerform seasonal decomposition analysis and obtain seasonal characteristics JJX tz , according to the observed trend GC qs and seasonal characteristics JJX tz FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk Fitting and modeling are performed to obtain the established time prediction model SJ mx ;

[0074] The implementation steps for extracting the most recent individual lifestyle habits, genetic factors, and environmental impact data to update the time prediction model are as follows:

[0075] Step ①: First, receive the data acquisition frequency insufficient command data PL bz Time prediction model SJ mx From the database SJ k Extract the individual lifestyle habits of the most recent stage from SH xg , Genetic factors ys 、Environmental impact yx data;

[0076] Step 2: SH your recent individual lifestyle habits xg , Genetic factors ys 、Environmental impact yx The data is based on the 70% training set XL j and 30% validation set YZ j Divide and then transform the time prediction model SJ mx Applied to training dataset XL sjj , and then the time prediction model SJ mx and training individual living habits SH xg , Genetic factors ys 、Environmental impact yx Match the features in the data and generate the prediction result YC jg ;

[0077] Step 3: Use the time prediction model SJ mx Generated prediction result YC jg With the training dataset XL sjj The actual observed value SJ in gcz Compare and compare the time prediction model SJ mx Output and actual observed value SJ gcz , and observe the time prediction model SJ mx How well it fits the training data?

[0078] Step ④: SJ based on time prediction modelmx Output and actual observed value SJ gcz The difference between the adjusted time prediction model SJ mx The parameters of the SJ prediction model are adjusted in time mx After the parameters are set, reapply the time prediction model SJ mx Based on the training data, a new round of time prediction model SJ is performed mx The output is compared with the actual observation value. This process is usually an iterative process until the time prediction model SJ mx Until the performance reaches a satisfactory level, it shows that the time prediction model SJ mx Update completed.

[0079] The evaluation and display unit 3 includes an indicator evaluation module 31 and a data display module 32;

[0080] The index evaluation module 31 is used to receive the time prediction model established in the data model unit 2 and the individual health status data analyzed in the status analysis module 13, and evaluate the individual health index according to the established time prediction model and the analyzed individual health status data, and obtain the evaluated individual health index data, which is used to evaluate the body mass index and cardiovascular health index data. Through the time prediction model and the analysis of health data, the individual's health status can be more comprehensively and systematically evaluated, including physical indicators, living habits, and disease risks, to provide more accurate health evaluation results;

[0081] The data display module 32 is used to receive the individual health index data evaluated in the index evaluation module 31 and display it to the outside world;

[0082] The data model unit 2 receives the individual health indicator data evaluated in the indicator evaluation module 31, and the updated time prediction model predicts the disease risk of the evaluated individual health indicator data and the extracted historical individual health data, predicts the individual's future cardiovascular disease, diabetes, and cancer, and obtains predicted disease risk data. Based on the individual's disease risk prediction, a personalized health management plan can be tailored for each person, including lifestyle intervention and regular check-ups, to improve the accuracy and effectiveness of health management.

[0083] The intervention suggestion unit 4 is used to receive the disease risk data predicted in the data model unit 2. The intervention suggestion unit 4 extracts historical individual health data from the database in the analysis and judgment module 11, and makes lifestyle intervention suggestions based on the predicted disease risk data and the environmental data, living habits, genetic factors and medical record data in the extracted historical individual health data, providing individuals with targeted diet adjustments, exercise plans, and sleep management. Through effective lifestyle intervention, the risk of diseases, including cardiovascular disease, diabetes, and obesity, can be reduced, providing individuals with a healthier lifestyle.

[0084] Usage process:

[0085] The analysis and judgment module 11 monitors and obtains individual health data, and obtains historical individual health data, and then analyzes the data quality based on the obtained individual health data, and uses the analyzed data quality and the set standard data quality to judge whether the data collection frequency is insufficient, and transmits the data collection frequency insufficient command data to the data fusion module 12. The data fusion module 12 extracts individual health data and historical individual health data from the database in the analysis and judgment module 11, and fuses the individual health data and the historical individual health data into multi-source data, and transmits the fused multi-source data to the status analysis module 13. The status analysis module 13 receives the analysis and judgment The data in module 11 or the multi-source data fused in the data fusion module 12, and the extracted individual health data or the fused multi-source data or the deduplicated individual health data are analyzed for individual health status, and the analyzed individual health status data are transmitted to the data model unit 2. The data model unit 2 establishes a time prediction model based on the analyzed individual health status data, the extracted individual health data and the extracted historical individual health data. When the data collection frequency is insufficient command data is received, the data model unit 2 extracts the recent individual living habits, genetic factors, and environmental impact data from the database in the analysis and judgment module 11 to update the time prediction model.

[0086] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A health monitoring time prediction system based on big data analysis, characterized by: It includes acquisition and analysis unit (1), data model unit (2), evaluation and presentation unit (3), and intervention suggestion unit (4); The acquisition and analysis unit (1) acquires individual health data and historical individual health data, performs data quality analysis based on the acquired individual health data and determines whether the data collection frequency is insufficient, and then performs individual health status analysis based on the acquired individual health data; The acquisition and analysis unit (1) comprises an analysis and judgment module (11), a data fusion module (12) and a situation analysis module (13); The analysis and judgment module (11) uses the monitoring device to monitor individual health data in real time and obtain individual health data, and simultaneously obtains historical individual health data, stores the obtained individual health data and the obtained historical individual health data in a database to establish a complete storage system, and then analyzes the data quality based on the obtained individual health data, and uses the analyzed data quality and the set standard data quality to determine whether the data collection frequency is insufficient; The specific judgment situations in the analysis and judgment module (11) include: Case ①: When the quality of the analyzed data is less than the set standard data quality, it means that the acquired individual health data has data missing, resulting in insufficient data collection frequency, and the insufficient data collection frequency command data is transmitted to the data fusion module (12); Case ②: When the quality of the analyzed data is greater than the set standard data quality, it indicates that the acquired individual health data has data duplication, and the acquired individual health data is deduplicated, and the deduplicated individual health data is transferred to the status analysis module (13); Case ③: When the quality of the analyzed data is consistent with the set standard data quality, it means that the acquired individual health data does not have data missing and data duplication, and the command data of no data missing and data duplication is transmitted to the status analysis module (13); Historical individual health data includes environmental data, lifestyle habits, genetic factors, social data, and medical records; The data fusion module (12) is used to receive the data collection frequency insufficient command data in the analysis and judgment module (11), the data fusion module (12) extracts the individual health data and the historical individual health data from the database in the analysis and judgment module (11), and fuses the individual health data and the historical individual health data into multi-source data; The condition analysis module (13) is used to receive the command data without data missing and data duplication or the individual health data without duplication or the multi-source data fused in the data fusion module (12) in the analysis and judgment module (11). When the command data without data missing and data duplication is received, the condition analysis module (13) extracts the individual health data from the database in the analysis and judgment module (11), and uses K-means in big data analysis to analyze the individual health status of the extracted individual health data or the fused multi-source data or the deduplicated individual health data; The implementation steps of using K-means in the condition analysis module (13) to analyze the individual health status of the extracted individual health data or the fused multi-source data or the deduplicated individual health data are as follows: Step 1: First collect the extracted individual health data GT jk or fused multi-source data DY sj Or deduplicated individual health data QC jk , and then extract the individual health data GT jk or fused multi-source data DY sj Or deduplicated individual health data QC jk Get feature T z ; Feature T z Including physiological index characteristic data SL zb , motion characteristic data YD tz , Sleep characteristic data SM tz , Eating habits characteristic data YS xg and weight characteristic data TZ tz ; Step 2: Based on the physiological index characteristic data SL zb , motion characteristic data YD tz , Sleep characteristic data SM tz , Eating habits characteristic data YS xg and weight characteristic data TZ tz Find out the features T that are more correlated with individual health status z ; Step ③: According to the characteristics T that are more correlated with individual health status z Determine the K value and select the K value as the initial cluster center JL zx , and then the characteristics T with higher correlation with individual health status are obtained z To each cluster center JL zx The distance between the individual and the health status of the individual is higher. z Assigned to the closest cluster center JL zx , update cluster center JL zx until the convergence condition is reached and the cluster JL is obtained c ; Step ④: According to cluster JL c The features combined with domain knowledge zs and experience y Each cluster c is explained, the individual health status represented by each cluster c is explained, and the analyzed individual health status data FX is obtained. jk ; The data model unit (2) is used to receive and acquire data in the analysis unit (1), and the data model unit (2) acquires individual health data and historical individual health data from the analysis and judgment module (11) and establishes a time prediction model and updates the time prediction model; The data model unit (2) is used to receive the individual health status data analyzed by the status analysis module (13) and the data collection frequency insufficient command data in the analysis and judgment module (11); the data model unit (2) extracts the individual health data and the historical individual health data from the database in the analysis and judgment module (11); the data model unit (2) uses time series analysis to establish a time prediction model based on the analyzed individual health status data, the extracted individual health data and the extracted historical individual health data; the prediction model is used to collect individual living habits, genetic factors, and environmental impact data; when the data collection frequency insufficient command data is received, the data model unit (2) extracts the individual living habits, genetic factors, and environmental impact data of the most recent stage from the database in the analysis and judgment module (11) to update the time prediction model; The implementation steps of building a time prediction model using time series analysis are as follows: Step 1: First collect and analyze the individual health status data FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk , and draw the analyzed individual health status data FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk Timing diagram SX over time t ; Step 2: Draw the timing diagram SX t FX, the individual health data for detection and analysis jk , extracted individual health data GT jk and extracted historical individual health data LS jk Is there any autocorrelation? Observe the individual health data of the analysis at a certain point in time. jk , extracted individual health data GT jk and extracted historical individual health data LS jk Whether the value is related to the previous time point; Step 3: Observe the timing diagram SX t Does the data show a clear upward or downward trend and draw an observed trend? qs , and then according to the timing diagram SX t Perform seasonal decomposition analysis and obtain seasonal characteristics JJX tz , according to the observed trend GC qs and seasonal characteristics JJX tz FX jk , extracted individual health data GT jk and extracted historical individual health data LS jk Fitting and modeling are performed to obtain the established time prediction model SJ mx ; The implementation steps for extracting the most recent individual lifestyle habits, genetic factors, and environmental impact data to update the time prediction model are as follows: Step ①: First, receive the data acquisition frequency insufficient command data PL b When z, the time prediction model SJ mx From the database SJ k Extract the most recent individual lifestyle habits from SH xg , Genetic factors ys 、Environmental impact yx data; Step 2: SH your recent individual lifestyle habits xg , Genetic factors ys 、Environmental impact yx The data is based on the 70% training set XL j and 30% validation set YZ j Divide and then transform the time prediction model SJ mx Applied to training dataset XL sjj , and then the time prediction model SJ mx and training individual living habits SH xg , Genetic factors ys 、Environmental impact yx Match the features in the data and generate the prediction result YC jg ; Step 3: Use the time prediction model SJ mx Generated prediction result YC jg With the training dataset XL sjj The actual observed value SJ in gcz Compare and contrast the time prediction model SJ mx Output and actual observed value SJ gcz , and observe the time prediction model SJ mx How well it fits the training data; Step ④: SJ based on time prediction model mx Output and actual observed value SJ gcz The difference between the adjusted time prediction model SJ mx The parameters of the SJ prediction model are adjusted in time mx After the parameters are set, reapply the time prediction model SJ mx Based on the training data, a new round of time prediction model SJ is performed mx The output is compared with the actual observation value. This process is usually an iterative process until the time prediction model SJ mx Until the performance reaches a satisfactory level, it shows that the time prediction model SJ mx Update completed; The evaluation and display unit (3) is used to receive the time prediction model established in the data model unit (2) and obtain the individual health status data analyzed in the analysis unit (1) and evaluate the individual health index, and then display the evaluated individual health index data to the outside world; The intervention suggestion unit (4) is used to receive the data in the data model unit (2) and obtain historical individual health data from the analysis and judgment module (11), and to make lifestyle intervention suggestions based on the data in the data model unit (2) and the obtained historical individual health data.

2. The health monitoring time prediction system based on big data analysis according to claim 1 is characterized by: The evaluation and display unit (3) comprises an indicator evaluation module (31) and a data display module (32); The index evaluation module (31) is used to receive the time prediction model established in the data model unit (2) and the individual health status data analyzed in the status analysis module (13), and evaluate the individual health index according to the established time prediction model and the analyzed individual health status data to obtain the evaluated individual health index data, and the evaluated individual health index data is used to evaluate the body mass index and cardiovascular health index data; The data display module (32) is used to receive the individual health index data evaluated in the index evaluation module (31) and display it to the outside world.

3. The health monitoring time prediction system based on big data analysis according to claim 2 is characterized by: The data model unit (2) receives the individual health index data evaluated in the index evaluation module (31), and the updated time prediction model predicts the disease risk of the evaluated individual health index data and the extracted historical individual health data, predicts the individual's future cardiovascular disease, diabetes, and cancer, and obtains predicted disease risk data.

4. The health monitoring time prediction system based on big data analysis according to claim 3 is characterized by: The intervention suggestion unit (4) is used to receive the disease risk data predicted by the data model unit (2), extract historical individual health data from the database in the analysis and judgment module (11), and make lifestyle intervention suggestions based on the predicted disease risk data and the environmental data, living habits, genetic factors and medical record data in the extracted historical individual health data.

Images