Health risk early warning management system and method based on multi-modal fusion

Abstract

The invention discloses a health risk early warning management system and method based on multi-modal fusion, relates to the technical field of health risk early warning, and discloses a multi-modal fusion module, a health risk user marking module, a health risk processing prediction module, a health risk early warning module and a health risk management module. A health risk processing prediction module, a health risk early warning module and a health risk management module are set, health risk processing of health risk users is simulated by constructing a health risk simulation processing model, a plurality of targeted health risk processing methods are generated, health risk processing difficulty is comprehensively evaluated, and a health risk management result is obtained. The health risk of the health risk user is subjected to graded early warning in combination with the prediction and processing difficulty of the health risk, so that a doctor can accurately know the health risk of the health risk user, and the health risk processing method is dynamically adjusted according to the actual condition of the health risk user.

Description

Technical Field

[0001] The present invention relates to the technical field of health risk early warning, and more specifically, to a health risk early warning management system and method based on multimodal fusion. Background Art

[0002] With the continuous advancement of modern medical technology, people are increasingly concerned about the management and prevention of personal health status. Traditional health management methods mainly rely on the monitoring of a single health indicator, such as heart rate, blood pressure or blood sugar, but the monitoring of these single indicators is often difficult to fully reflect the overall health status and potential risks of an individual. Therefore, health management systems that can monitor multimodal data have gradually appeared on the market. These systems provide users with a more comprehensive health assessment by collecting and analyzing a variety of health data of users.

[0003] However, although multi-indicator monitoring has improved the comprehensiveness of health management to a certain extent, most existing systems are still at the data collection level and lack in-depth integration and analysis of these multimodal data. In addition, even if some systems can provide some health advice or warnings, these advice often lack personalization and pertinence, making it difficult to meet the actual needs of different users.

[0004] In terms of health risk warning, existing methods usually rely on statistical analysis of historical data. Although they can predict certain health risks to a certain extent, the prediction accuracy and practicality still need to be improved. Summary of the invention

[0005] In view of the deficiencies in the prior art, the object of the present invention is to provide a health risk warning management system and method based on multimodal fusion.

[0006] To achieve the above object, the present invention provides the following technical solutions: A health risk early warning management system based on multimodal fusion, including a multimodal fusion module, a health risk user marking module, a health risk processing prediction module, a health risk early warning module, and a health risk management module; The multimodal fusion module undergoes a health monitoring period , obtain the user's real modality fusion feature set; The health risk user marking module obtains a health risk depth prediction value based on the user's real modality fusion feature set, and determines whether to mark the user as a health risk user based on a comparison result between the health risk depth prediction value and the health risk depth prediction threshold value; The health risk processing prediction module is used to construct a health risk simulation processing model for health risk users. The health risk simulation processing model performs j simulation processings, obtains the health risk processing difficulty value of each simulation processing, and then obtains the health risk processing prediction value of the health risk user; The health risk warning module is used to obtain the health risk warning value of the health risk user and determine the health risk warning level of the health risk user based on the health risk warning value; The health risk management module is used to sort the health risk processing difficulty values ​​of all simulated processing from small to large, mark the health risk processing method corresponding to the simulated processing ranked first as the preferred processing method, and use the preferred processing method to perform health risk processing on health risk users. During the health risk processing process, if the preferred processing method is not suitable for the health risk user, the health risk processing method corresponding to the next simulated processing ranked first is marked as the preferred processing method.

[0007] Furthermore, after each health monitoring period , obtain the user's real modality fusion feature set, specifically: collect health monitoring time The real data of each modal data of the user is obtained, and features are extracted from the real data of each modal data to obtain the real features of each modal data, and the real features of each modal data are combined into a real modal fusion feature set.

[0008] Furthermore, a deep prediction value of health risk is obtained based on the user's real modal fusion feature set, specifically: a modal fusion feature prediction model is obtained, the real modal fusion feature set is used as input data of the modal fusion feature prediction model, and the modal fusion feature prediction model outputs the deep prediction value of health risk.

[0009] Furthermore, a health risk simulation processing model for health risk users is constructed, specifically: an ideal modal fusion feature set of health risk users is obtained, historical data of each modal data of health risk users are simultaneously obtained, and a health risk simulation processing model is constructed based on the real modal fusion feature set of health risk users, the ideal modal fusion feature set and historical data of each modal data.

[0010] Furthermore, the difficulty value of the health risk treatment for the simulation is obtained as follows: Set the duration of the simulation period During the simulation process of the health risk simulation model, the duration of each simulation period is , obtain a health risk depth prediction value. When the health risk simulation processing model completes the simulation processing, all health risk depth prediction values ​​are sorted in the order of acquisition, and the difference between the previous health risk depth prediction value and the next health risk depth prediction value after sorting is calculated to obtain the health risk processing value. The health risk processing threshold is set. When the health risk processing value is greater than or equal to the health risk processing threshold, the number of effective health risk processing times is increased once, and the number of effective health risk processing times is marked as Lddw. The total number of simulation time periods is marked as Pedg. The health risk continuation times Kgat are obtained. The formula is used The health risk treatment difficulty value Ustz of this simulation is obtained, where pa is the effective health risk treatment coefficient, pb is the simulation period quantity coefficient, and pc is the health risk continuation coefficient.

[0011] Furthermore, the health risk continuation number Kgat is obtained as follows: sum the two adjacent health risk depth prediction values ​​after sorting to obtain the health risk prediction continuation value, set the health risk prediction continuation limit, and when the health risk prediction continuation value is greater than or equal to the health risk prediction continuation limit, increase the health risk continuation number by one, and mark the health risk continuation number as Kgat.

[0012] Furthermore, the health risk treatment prediction value of the health risk user is obtained as follows: a high value of health risk treatment difficulty and a low value of health risk treatment difficulty are set. When the health risk treatment difficulty value of the simulated treatment is greater than or equal to the high value of health risk treatment difficulty, the number of health risk difficulty treatments is increased once. When the health risk treatment difficulty value of the simulated treatment is less than or equal to the low value of health risk treatment difficulty, the number of simple health risk treatments is increased once. When the health risk treatment difficulty value of the simulated treatment is between the high value of health risk treatment difficulty and the low value of health risk treatment difficulty, the health risk treatment difficulty value is marked as a regular treatment difficulty value. All regular treatment difficulty values ​​are summed and averaged to obtain the regular treatment difficulty mean Zrwp. The number of health risk difficulty treatments is marked as Szby, and the number of simple health risk treatments is marked as Fged. The formula is used. The health risk treatment prediction value Bht is obtained.

[0013] Furthermore, the health risk warning value of the health risk user is obtained as follows: the health risk depth prediction value Arp and the health risk processing difficulty value Bht of the health risk user are obtained, and the formula is used The health risk warning value gdz of the health risk user is obtained, where y1 is the health risk depth prediction coefficient and y2 is the health risk processing difficulty coefficient.

[0014] Furthermore, the health risk early warning management method based on multimodal fusion includes the following steps: Step 1: After each health monitoring period , obtain the user's real modality fusion feature set; Step 2: Obtain a health risk depth prediction value based on the user's real modality fusion feature set, and determine whether to mark the user as a health risk user based on the comparison result between the health risk depth prediction value and the health risk depth prediction threshold; Step 3: Construct a health risk simulation processing model for health risk users. The health risk simulation processing model performs j simulation processings, obtains the health risk processing difficulty value for each simulation processing, and then obtains the health risk processing prediction value for the health risk user; Step 4: Obtain the health risk warning value of the health risk user, and determine the health risk warning level of the health risk user based on the health risk warning value; Step 5: Sort the health risk processing difficulty values ​​of all simulated processing from small to large, mark the health risk processing method corresponding to the simulated processing that is ranked first as the preferred processing method, and use the preferred processing method to process the health risk of the health risk user. During the health risk processing process, if the preferred processing method is not suitable for the health risk user, mark the health risk processing method corresponding to the next simulated processing in the ranking as the preferred processing method.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. The method of the present invention combines the prediction and processing difficulty of health risks to provide graded warnings for the health risks of health risk users, so that doctors can accurately understand the health risks of health risk users and dynamically adjust the health risk processing method according to the actual situation of health risk users; 2. Set up a multimodal fusion module and a health risk user marking module to collect and fuse the user's multimodal health risk data, accurately predict whether the user has health risks, set up a health risk processing prediction module, a health risk warning module and a health risk management module, and simulate the health risk processing of health risk users by building a health risk simulation processing model, generate multiple targeted health risk processing methods, and comprehensively evaluate the difficulty of health risk processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a method flow chart of the health risk early warning management method based on multimodal fusion; Figure 2 This is a system module diagram of a health risk early warning management system based on multimodal fusion; Figure 3 Figure 2. Methodology for constructing a model for health risk simulation treatment. DETAILED DESCRIPTION

[0017] Example 1: Reference Figure 1 , a health risk early warning management method based on multimodal fusion includes the following steps: Step 1: After each health monitoring period , obtain the user's real modality fusion feature set; Step 2: Obtain a health risk depth prediction value based on the user's real modality fusion feature set, and determine whether to mark the user as a health risk user based on the comparison result between the health risk depth prediction value and the health risk depth prediction threshold; Step 3: Construct a health risk simulation processing model for health risk users. The health risk simulation processing model performs j simulation processings, obtains the health risk processing difficulty value for each simulation processing, and then obtains the health risk processing prediction value for the health risk user; Step 4: Obtain the health risk warning value of the health risk user, and determine the health risk warning level of the health risk user based on the health risk warning value; Step 5: Sort the health risk processing difficulty values ​​of all simulated processing from small to large, mark the health risk processing method corresponding to the simulated processing that is ranked first as the preferred processing method, and use the preferred processing method to process the health risk of the health risk user. During the health risk processing process, if the preferred processing method is not suitable for the health risk user, mark the health risk processing method corresponding to the next simulated processing in the ranking as the preferred processing method.

[0018] The above method combines the prediction of health risks and the difficulty of handling them to provide graded warnings on the health risks of health risk users, making it easier for doctors to accurately understand the health risks of health risk users and dynamically adjust the health risk handling methods according to the actual conditions of the health risk users.

[0019] Example 2: Reference Figures 2 to 3 The health risk early warning management system based on multimodal fusion includes a multimodal fusion module, a health risk user marking module, a health risk processing prediction module, a health risk early warning module, and a health risk management module.

[0020] Multimodal fusion module: Setting health monitoring duration ( The specific value of the health monitoring duration. During system operation, the health monitoring duration Infinite loop, The value is adjusted according to the system requirements and the actual situation of the user). , collect health monitoring time The real data of each modal data of the user (the user's modal data includes heart rate data, blood pressure data, blood sugar data, etc., and each modal data is collected through corresponding sensors and wearable devices), the real data of each modal data is feature extracted to obtain the real features of each modal data, and the real features of each modal data are combined into a real modal fusion feature set.

[0021] Health risk user marking module: obtain the modal fusion feature prediction model, use the real modal fusion feature set as the input data of the modal fusion feature prediction model, and the modal fusion feature prediction model outputs the health risk depth prediction value. Set the health risk depth prediction threshold (the health risk depth prediction threshold is the system preset value). When the health risk depth prediction value is greater than or equal to the health risk depth prediction threshold, the user is marked as a health risk user. When the health risk depth prediction value is greater than or equal to the health risk depth prediction threshold, no processing is performed.

[0022] The construction process of the modal fusion feature prediction model is as follows: collect i groups of real modal fusion feature sets, build a deep learning model, use the real modal fusion feature sets as training data for the deep learning model, and assign a health risk depth prediction value to each training data. The range of the health risk depth prediction value is (1.0~2.9). The larger the health risk depth prediction value, the longer the user is predicted to spend in the next health monitoring. The greater the possibility of health risk, the smaller the health risk depth prediction value, and the predicted duration of the next health monitoring of the user. The smaller the possibility of health risks, the training data is divided into training set, validation set and test set according to the set ratio of 3:1:1. The training set, validation set and test set are trained. After the training, the modal fusion feature prediction model is constructed.

[0023] A multimodal fusion module and a health risk user marking module are set up to collect and fuse the user's multimodal health risk data to accurately predict whether the user has health risks.

[0024] Health risk processing prediction module: obtain the ideal modal fusion feature set of health risk users (the ideal modal fusion feature set is a combination of ideal features of each modal data, and the ideal features are the features of each modal data of the user in a state of no health risk), and synchronously obtain the historical data of each modal data of the health risk user (the historical data is the real data of each modal data obtained from the health risk user before the current time of the system), and build a health risk simulation processing model based on the real modal fusion feature set of the health risk user, the ideal modal fusion feature set and the historical data of each modal data. The health risk simulation processing model performs j simulation processings (each simulation processing uses a health risk processing method, and after each simulation processing, the user model in the health risk simulation processing model satisfies the features of each modal data in the ideal modal fusion feature set), obtains the health risk processing difficulty value of each simulation processing, and then obtains the health risk processing prediction value of the health risk user.

[0025] A health risk simulation processing model is constructed based on the real modal fusion feature set of health risk users, the ideal modal fusion feature set and the historical data of each modal data. Specifically, MATLAB / Simulink simulation software is selected, and a user model is created in the simulation software based on the real modal fusion feature set of health risk users of safety risk equipment and the historical data of each modal data, and physiological parameters are set for the user model, and the health risk processing direction of the user model is determined based on the ideal modal fusion feature set of health risk users, and a health risk simulation processing model is constructed. The health risk simulation processing model generates a health risk processing method according to the health risk processing direction, and the health risk simulation processing model uses the health risk processing method to perform health risk processing on the user model in the health risk simulation processing model.

[0026] The difficulty value of health risk treatment for simulation is obtained as follows: Set the duration of the simulation period During the simulation process of the health risk simulation model, the duration of each simulation period is , obtain a health risk depth prediction value. When the health risk simulation processing model completes the simulation processing, all health risk depth prediction values ​​are sorted in the order of acquisition. The difference between the previous health risk depth prediction value and the next health risk depth prediction value after sorting is calculated to obtain the health risk processing value. The health risk processing threshold is set (the health risk processing threshold is the system preset value). When the health risk processing value is greater than or equal to the health risk processing threshold, the number of effective health risk processing times is increased once. When the health risk processing value is less than the health risk processing threshold, no processing is performed. The number of effective health risk processing times is marked as Lddw, and the total number of simulation time periods is marked as Pedg. The two adjacent health risk depth prediction values ​​after sorting are summed to obtain the health risk prediction continuation value. The health risk prediction continuation limit is set (the health risk prediction continuation limit is the system preset value). When the health risk prediction continuation value is greater than or equal to the health risk prediction continuation limit, the number of health risk continuations is increased once. When the health risk prediction continuation value is less than the health risk prediction continuation limit, no processing is performed. The number of health risk continuations is marked as Kgat. Using the formula The health risk treatment difficulty value Ustz of this simulation is obtained, where pa is the effective health risk treatment coefficient, pb is the simulation period quantity coefficient, pc is the health risk continuation coefficient, the value of pa is 0.75, the value of pb is 0.41, and the value of pc is 0.34.

[0027] The method for obtaining the health risk processing prediction value of the health risk user is as follows: set the high value of the health risk processing difficulty and the low value of the health risk processing difficulty (the high value of the health risk processing difficulty is greater than the low value of the health risk processing difficulty, and the high value and the low value of the health risk processing difficulty are both system preset values). When the simulated health risk processing difficulty value is greater than or equal to the high value of the health risk processing difficulty, the number of health risk difficulty processing times is increased once. When the simulated health risk processing difficulty value is less than or equal to the low value of the health risk processing difficulty, the number of simple health risk processing times is increased once. When the simulated health risk processing difficulty value is between the high value of the health risk processing difficulty and the low value of the health risk processing difficulty, the health risk processing difficulty value is marked as a regular processing difficulty value. All regular processing difficulty values ​​are summed and averaged to obtain the regular processing difficulty mean Zrwp. The number of health risk difficulty processing times is marked as Szby, and the number of simple health risk processing times is marked as Fged. Use the formula The health risk treatment prediction value Bht is obtained.

[0028] Health risk warning module: obtains the health risk warning value of the health risk user, and determines the health risk warning level of the health risk user based on the health risk warning value.

[0029] The health risk warning value of the health risk user is obtained as follows: obtain the health risk depth prediction value Arp and the health risk processing difficulty value Bht of the health risk user, and use the formula The health risk warning value gdz of the health risk user is obtained, where y1 is the health risk depth prediction coefficient, y2 is the health risk processing difficulty coefficient, the value of y1 is 0.91, and the value of y2 is 0.83.

[0030] The health risk warning level of the health risk user is determined based on the health risk warning value, specifically: the health risk warning value gdz of the health risk user, the range of each health risk warning value corresponds to a health risk warning level, the range of the health risk warning value is (0, gd1], (gd1, gd2], …, (gdz-1, gdz], the health risk warning levels include health risk warning level 1, health risk warning level 2, …, health risk warning level z-1, health risk warning level z, when the health risk warning value gdz∈(0, gd1], the health risk user is health risk warning level 1, the health risk warning intensity of health risk warning level 1 < the health risk warning intensity of health risk warning level 2 < … < the health risk warning intensity of health risk warning level z-1 < the health risk warning intensity of health risk warning level z, the doctor can accurately understand the health risk of the health risk user according to the health risk warning level.

[0031] Health risk management module: Sort the difficulty values ​​of all simulated health risk treatments from small to large, mark the health risk treatment method corresponding to the simulated treatment that ranks first as the preferred treatment method, and use the preferred treatment method to treat health risks for health risk users. During the health risk treatment process, if the preferred treatment method is not suitable for the health risk user, mark the health risk treatment method corresponding to the next simulated treatment in the order as the preferred treatment method.

[0032] A health risk treatment prediction module, a health risk early warning module and a health risk management module are set up. By constructing a health risk simulation processing model, the health risk treatment of health risk users is simulated, and multiple targeted health risk treatment methods are generated. The difficulty of health risk treatment is comprehensively evaluated. The health risks of health risk users are graded and early warnings are given based on the prediction and treatment difficulty of health risks. This facilitates doctors to accurately understand the health risks of health risk users and dynamically adjust the health risk treatment methods according to the actual situation of health risk users.

[0033] The above formulas are all dimensionless and numerical calculations are performed, and the preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0034] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0035] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0036] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0037] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0038] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0039] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0040] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A health risk early warning management system based on multimodal fusion, characterized in that: It includes multimodal fusion module, health risk user marking module, health risk processing prediction module, health risk early warning module and health risk management module; The multimodal fusion module undergoes a health monitoring period , obtain the user's real modality fusion feature set; The health risk user marking module obtains a health risk depth prediction value based on the user's real modality fusion feature set, and determines whether to mark the user as a health risk user based on a comparison result between the health risk depth prediction value and the health risk depth prediction threshold value; The health risk processing prediction module is used to construct a health risk simulation processing model for health risk users. The health risk simulation processing model performs j simulation processings, obtains the health risk processing difficulty value of each simulation processing, and then obtains the health risk processing prediction value of the health risk user; The health risk warning module is used to obtain the health risk warning value of the health risk user and determine the health risk warning level of the health risk user based on the health risk warning value; The health risk management module is used to sort the health risk processing difficulty values ​​of all simulated processing from small to large, mark the health risk processing method corresponding to the simulated processing ranked first as the preferred processing method, and use the preferred processing method to perform health risk processing on health risk users. During the health risk processing process, if the preferred processing method is not suitable for the health risk user, the health risk processing method corresponding to the next simulated processing ranked first is marked as the preferred processing method.

2. The health risk early warning management system based on multimodal fusion according to claim 1 is characterized in that: After each health monitoring period , obtain the user's real modality fusion feature set, specifically: collect health monitoring time The real data of each modal data of the user is obtained, and features are extracted from the real data of each modal data to obtain the real features of each modal data, and the real features of each modal data are combined into a real modal fusion feature set.

3. The health risk early warning management system based on multimodal fusion according to claim 1 is characterized in that: A deep prediction value of health risk is obtained based on the user's real modal fusion feature set, specifically: a modal fusion feature prediction model is obtained, the real modal fusion feature set is used as input data of the modal fusion feature prediction model, and the modal fusion feature prediction model outputs the deep prediction value of health risk.

4. The health risk early warning management system based on multimodal fusion according to claim 1, characterized in that: Construct a health risk simulation processing model for health risk users, specifically: obtain the ideal modal fusion feature set of health risk users, synchronously obtain the historical data of each modal data of the health risk users, and construct a health risk simulation processing model based on the real modal fusion feature set of the health risk users, the ideal modal fusion feature set and the historical data of each modal data.

5. The health risk early warning management system based on multimodal fusion according to claim 1 is characterized in that: The difficulty value of health risk treatment for simulation is obtained as follows: Set the duration of the simulation period During the simulation process of the health risk simulation model, the duration of each simulation period is , obtain a health risk depth prediction value. When the health risk simulation processing model completes the simulation processing, all health risk depth prediction values ​​are sorted in the order of acquisition, and the difference between the previous health risk depth prediction value and the next health risk depth prediction value after sorting is calculated to obtain the health risk processing value. The health risk processing threshold is set. When the health risk processing value is greater than or equal to the health risk processing threshold, the number of effective health risk processing times is increased once, and the number of effective health risk processing times is marked as Lddw. The total number of simulation time periods is marked as Pedg. The health risk continuation times Kgat are obtained. The formula is used The health risk treatment difficulty value Ustz of this simulation is obtained, where pa is the effective health risk treatment coefficient, pb is the simulation period quantity coefficient, and pc is the health risk continuation coefficient.

6. The health risk early warning management system based on multimodal fusion according to claim 5 is characterized in that: The method for obtaining the health risk continuation number Kgat is as follows: sum the two adjacent health risk depth prediction values ​​after sorting to obtain the health risk prediction continuation value, set the health risk prediction continuation limit, and when the health risk prediction continuation value is greater than or equal to the health risk prediction continuation limit, increase the health risk continuation number by one, and mark the health risk continuation number as Kgat.

7. The health risk early warning management system based on multimodal fusion according to claim 1, characterized in that: The method for obtaining the health risk treatment prediction value of the health risk user is as follows: set the high value of the health risk treatment difficulty and the low value of the health risk treatment difficulty. When the health risk treatment difficulty value of the simulated treatment is greater than or equal to the high value of the health risk treatment difficulty, increase the number of health risk difficulty treatments by one time. When the health risk treatment difficulty value of the simulated treatment is less than or equal to the low value of the health risk treatment difficulty, increase the number of simple health risk treatments by one time. When the health risk treatment difficulty value of the simulated treatment is between the high value of the health risk treatment difficulty and the low value of the health risk treatment difficulty, mark the health risk treatment difficulty value as a regular treatment difficulty value, sum up all the regular treatment difficulty values ​​and take the average to obtain the regular treatment difficulty mean Zrwp, mark the health risk difficulty treatment times as Szby, mark the health risk simple treatment times as Fged, and use the formula The health risk treatment prediction value Bht is obtained.

8. The health risk early warning management system based on multimodal fusion according to claim 1 is characterized in that: The health risk warning value of the health risk user is obtained as follows: obtain the health risk depth prediction value Arp and the health risk processing difficulty value Bht of the health risk user, and use the formula The health risk warning value gdz of the health risk user is obtained, where y1 is the health risk depth prediction coefficient and y2 is the health risk processing difficulty coefficient.

9. A health risk early warning management method based on multimodal fusion, applied to a health risk early warning management system based on multimodal fusion according to any one of claims 1 to 8, characterized in that: The steps include: Step 1: After each health monitoring period , obtain the user's real modality fusion feature set; Step 2: Obtain a health risk depth prediction value based on the user's real modality fusion feature set, and determine whether to mark the user as a health risk user based on the comparison result between the health risk depth prediction value and the health risk depth prediction threshold; Step 3: Construct a health risk simulation processing model for health risk users. The health risk simulation processing model performs j simulation processings, obtains the health risk processing difficulty value for each simulation processing, and then obtains the health risk processing prediction value for the health risk user; Step 4: Obtain the health risk warning value of the health risk user, and determine the health risk warning level of the health risk user based on the health risk warning value; Step 5: Sort the health risk processing difficulty values ​​of all simulated processing from small to large, mark the health risk processing method corresponding to the simulated processing that is ranked first as the preferred processing method, and use the preferred processing method to process the health risk of the health risk user. During the health risk processing process, if the preferred processing method is not suitable for the health risk user, mark the health risk processing method corresponding to the next simulated processing in the ranking as the preferred processing method.

Images