An artificial intelligence big data wisdom health management method and system

Abstract

The application relates to the technical field of intelligent health management, and particularly discloses an artificial intelligence big data intelligent health management method and system, which comprises the following steps: acquiring the body parameters of a user based on an intelligent device, constructing a body portrait of the user according to the body parameters; acquiring the behavior data of the user containing a time span, acquiring the change information of the body portrait according to the time span, and training a mapping model of the behavior data to the change information; comparing the latest body portrait and the latest mapping model of different users, and clustering the users; for any type of user, when receiving the physical examination data uploaded by any user, generating prompt information according to the physical examination data, and broadcasting to the same type of users; the data acquired by the intelligent device is used for clustering the users, highly similar users are classified into one type, when any user of the same type of users performs physical examination, the physical examination data is shared, and the physical examination frequency of each user is indirectly improved.

Description

Technical Field

[0001] This invention relates to the field of smart health management technology, specifically to an artificial intelligence big data smart health management method and system. Background Technology

[0002] With social progress and technological development, more and more people will use various smart devices to collect their own body data, thereby assessing their physical condition, enabling them to know their own status and ensure their health.

[0003] Existing home health management devices are relatively inexpensive and lack precision. While they can achieve some results, the effects are minimal, and they are mostly used for weight monitoring, making it difficult to meet the needs of health management. Summary of the Invention

[0004] The purpose of this invention is to provide an artificial intelligence big data-based smart health management method and system to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] An artificial intelligence big data-driven smart health management method and system, the method comprising:

[0007] The user's body parameters are acquired using smart devices, and a body profile of the user is constructed based on the body parameters.

[0008] Acquire user behavior data spanning a time span, obtain information on changes in body profiles based on the time span, and train a mapping model from behavior data to change information;

[0009] By comparing the latest physical profiles and mapping models of different users, users are clustered.

[0010] For any type of user, when receiving physical examination data uploaded by any user, a prompt message is generated based on the physical examination data and broadcast to users of the same type.

[0011] As a further aspect of the present invention: the step of acquiring the user's body parameters based on a smart device and constructing a user's body profile based on the body parameters includes:

[0012] Based on a smart body fat scale, the user's basic parameters are obtained;

[0013] Based on the aforementioned basic parameters, the system iterates through and matches body models in a preset body model library.

[0014] The display parameters of the body model are determined based on the changes in user parameters acquired in real time by smart wearable devices.

[0015] The display parameters will be inserted into the body model to obtain the user's body profile.

[0016] The variable parameters include heart rate, exercise volume, and blood oxygen saturation. Heart rate, exercise volume, and blood oxygen saturation are all converted to the range of 0 to 255 and used as H value, S value, and V value, respectively, to obtain the display parameters.

[0017] As a further aspect of the present invention: the steps of acquiring user behavioral data with a time span, acquiring changes in body profiles based on the time span, and training a mapping model from behavioral data to changes in information include:

[0018] Based on real-time acquisition of user change parameters by smart wearable devices, the change parameters are identified to determine the user's behavior type at each time point;

[0019] Within a time period, the behavior types at various moments are merged to obtain the behavior data for each time period;

[0020] Obtain information on changes in body profiles at each time period, use behavioral data from the same time period as features, and use information on changes in body profiles as labels to construct samples;

[0021] Collect a predetermined number of samples, construct a sample set, and divide the sample set into a training set and a test set according to a predetermined ratio to train the neural network model;

[0022] When the error rate of the neural network model is less than a preset error rate threshold, the neural network model is output as a mapping model from behavioral data to change information.

[0023] As a further aspect of the present invention: the step of clustering users by comparing the latest body profiles and the latest mapping models of different users includes:

[0024] Obtain the body profile and mapping model of each user at the current moment;

[0025] Compare the body profiles and mapping models of different users, and calculate the profile distance and model distance;

[0026] Users are clustered based on the portrait distance and model distance.

[0027] As a further aspect of the present invention: the step of clustering users based on the profile distance and the model distance includes:

[0028] For any two users, read the body profiles of the two users, calibrate the origin of the two body profiles, and calculate the distance between the profiles;

[0029] Read the mapping models of two users, randomly determine a preset number of test data, input the test data into the two mapping models, obtain the output, and calculate the model distance based on the output;

[0030] User similarity is calculated based on image distance and model distance, and the users are clustered using the optics clustering algorithm; the similarity is inversely proportional to both image distance and model distance.

[0031] The calculation process for the image distance is as follows:

[0032] d1 represents the image distance; N, M, and K represent the dimensions of the space where the body model is located; A(i,j,k) and B(i,j,k) represent the parameters at point (i,j,k) in the two body models, respectively. A(i,j,k) and B(i,j,k) are both arrays used to represent the display parameters. When a body model does not display parameters at point (i,j,k), an array with all zero elements is used as a substitute array; dis{A(i,j,k),B(i,j,k)} represents the array distance between the two arrays; α is a preset adjustment coefficient.

[0033] As a further aspect of the present invention: the step of generating a prompt message based on the physical examination data and broadcasting it to users of the same type when receiving physical examination data uploaded by any user of any type includes:

[0034] Regularly send requests to various users to obtain physical examination data;

[0035] When any user uploads physical examination data, the data is transcoded to obtain digital physical examination data.

[0036] Locate abnormal data in digital health checkup data and insert a prompt box into the abnormal data;

[0037] Broadcast digital health check data containing prompts to similar users.

[0038] The present invention also provides an artificial intelligence big data smart health management system, the system comprising:

[0039] The body profile construction module is used to acquire the user's body parameters based on smart devices and construct the user's body profile based on the body parameters.

[0040] The mapping model training module is used to acquire users' behavioral data with a time span, obtain information on changes in body profiles based on the time span, and train a mapping model from behavioral data to change information.

[0041] The user clustering module is used to cluster users by comparing their latest body profiles and latest mapping models at different times.

[0042] The physical examination data sharing module is used to generate a prompt message based on the physical examination data uploaded by any user of any type, and broadcast it to users of the same type.

[0043] As a further aspect of the present invention: the body portrait construction module includes:

[0044] The basic parameter acquisition unit is used to acquire the user's basic parameters based on the smart body fat scale.

[0045] A body model matching unit is used to traverse and match body models in a preset body model library based on the basic parameters.

[0046] The variable parameter determination unit is used to acquire the user's variable parameters in real time based on the smart wearable device, and determine the display parameters of the body model based on the variable parameters.

[0047] The display parameter application unit is used to insert display parameters into the body model to obtain the user's body profile.

[0048] The variable parameters include heart rate, exercise volume, and blood oxygen saturation. Heart rate, exercise volume, and blood oxygen saturation are all converted to the range of 0 to 255 and used as H value, S value, and V value, respectively, to obtain the display parameters.

[0049] As a further aspect of the present invention: the mapping model training module includes:

[0050] The variable parameter identification unit is used to acquire the user's variable parameters in real time based on the smart wearable device, identify the variable parameters, and determine the user's behavior type at each time.

[0051] The behavior type merging unit is used to merge the behavior types at various times within a time period to obtain the behavior data for each time period.

[0052] The sample creation unit is used to obtain the change information of the body profile at each time period, and construct samples by using the behavioral data of the same time period as features and the change information of the body profile as labels.

[0053] The model training unit is used to count a preset number of samples, construct a sample set, and divide the sample set into a training set and a test set according to a preset ratio to train the neural network model.

[0054] The model output unit is used to output the neural network model as a mapping model from behavioral data to change information when the error rate of the neural network model is less than a preset error rate threshold.

[0055] As a further aspect of the present invention: the user clustering module includes:

[0056] The parameter acquisition unit is used to acquire the body profile and mapping model of each user at the current moment;

[0057] The user comparison unit is used to compare the body profiles and mapping models of different users and calculate the profile distance and model distance.

[0058] The clustering execution unit is used to cluster users based on the profile distance and the model distance.

[0059] Compared with the prior art, the beneficial effects of the present invention are: the present invention clusters users by acquiring data from smart devices, grouping highly similar users into one category. When any user in the same category undergoes a physical examination, the examination data is shared. When there are a large number of users, there will be physical examinations almost every day, which indirectly increases the frequency of physical examinations for each user and meets the health management needs of users. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention.

[0061] Figure 1 The overall flowchart of the AI ​​big data-driven smart health management method is shown.

[0062] Figure 2 The diagram shows the structure of the AI-powered big data-driven smart health management system. Detailed Implementation

[0063] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0064] Figure 1 This is a flowchart illustrating the overall process of an artificial intelligence big data-based smart health management method and system. In this embodiment of the invention, an artificial intelligence big data-based smart health management method includes:

[0065] Step S100: Obtain the user's body parameters based on the smart device, and construct the user's body profile based on the body parameters;

[0066] The intelligent devices in this invention include an intelligent body fat scale and wearable devices. Existing intelligent body fat scales can measure many body parameters, and many companies offer very affordable intelligent body fat scales, which are commonly used in many households. The data that body fat scales can measure includes: weight, BMI, body fat percentage, muscle mass, water percentage, basal metabolic rate (BMR), visceral fat level, bone mass, and body age, etc. The principle of obtaining these parameters is BIA (bioelectrical impedance analysis) technology (measuring body composition through foot electrodes). Its accuracy is not very high, but it is sufficient for the application scenario of this application. In addition, wearable devices refer to some smart bracelets that can monitor heart rate, steps, and blood oxygen saturation, etc. Body fat scales generally measure once at regular intervals, such as every few days, while wearable devices can monitor in real time, meaning every few minutes or even every few seconds. The combined monitoring data from the body fat scale and wearable devices are collectively referred to as body parameters. By analyzing and applying these body parameters, a user's body profile can be constructed. The body profile is a three-dimensional model of the user containing display parameters.

[0067] Step S200: Obtain user behavior data with a time span, obtain body profile change information based on the time span, and train a mapping model from behavior data to change information;

[0068] Behavioral data is a broader concept used to characterize user behavior, such as walking, standing, sitting, lying down, and running. Each behavior is a behavior over a period of time. For example, if a user is asleep from 0:00 to 8:00, the behavioral data with this time span is the sleep behavior from 0:00 to 8:00. At the same time, the changes in the body profile from 0:00 to 8:00 are the change information mentioned above. Using the behavioral data with the time span as the independent variable and the changes in the body profile as the dependent variable, a neural network model is trained to map the independent variables to the dependent variables. After training, a mapping model from behavioral data to change information is obtained.

[0069] Step S300: Compare the latest body profiles and the latest mapping models of different users to cluster the users;

[0070] For any given user, their latest body profile represents their latest state, and the latest mapping model represents the impact of their behavior on their body. Both of these can reflect the user's true state in a very comprehensive way. By comparing the latest body profiles and latest mapping models of different users, the differences between the two users are calculated based on the comparison results. Then, users are clustered, and users of the same type are grouped into one category.

[0071] It should be noted that the users mentioned in this application refer to users who have already registered with the subject of this method. During the registration phase, the subject of this method will send a permission acquisition request to obtain the permissions granted by the user. Only in this way can the subject of this method be able to query which category the user belongs to, and only then can it send physical examination data to the corresponding user in the subsequent physical examination data sharing process.

[0072] Step S400: For any type of user, when receiving physical examination data uploaded by any user, generate a prompt message based on the physical examination data and broadcast it to users of the same type;

[0073] Users within the same category have highly similar physical conditions and implicitly similar behavioral responses. When any user in any category undergoes a physical examination and uploads the data to the execution entity of this method, the execution entity can share the data with other users in the same category. However, this process requires explicit permission from the user. In practice, authorization can be obtained once during registration, and subsequent sharing can proceed directly. While this method is convenient, physical examination data is, after all, the user's personal data. Therefore, this application employs a multi-question approach, sending a permission grant request to the user again before each sharing. This method is not very convenient and may even be cumbersome for users, but it provides very clear authorization and is essentially for the user's benefit.

[0074] In one example of the technical solution of this invention, after a user registers on the execution subject of this method, they upload detection data in real time. The detection data is acquired by a smart device. The execution subject of this method categorizes users based on the detection data, grouping highly similar users into one category. At this time, when the user undergoes a highly accurate physical examination, the examination report will be shared with users of the same category. For an individual user, a physical examination may be once a year, but for a category of users, when there are many users, the probability of no one undergoing a physical examination on a particular day is actually very low. This also means that any user in the same category can receive a physical examination report almost every day, and the physical examination report received is from users who are very similar to themselves, enabling users to have a more accurate understanding of their own status.

[0075] As a preferred embodiment of the technical solution of the present invention, the step of acquiring the user's body parameters based on a smart device and constructing a user's body profile based on the body parameters includes:

[0076] Based on a smart body fat scale, the user's basic parameters are obtained;

[0077] Based on the aforementioned basic parameters, the system iterates through and matches body models in a preset body model library.

[0078] The display parameters of the body model are determined based on the changes in user parameters acquired in real time by smart wearable devices.

[0079] The display parameters will be inserted into the body model to obtain the user's body profile.

[0080] The system uses a smart body fat scale to acquire the user's basic parameters, including weight, height, body fat percentage, muscle mass, and bone mass. Fewer parameters can also be acquired. Based on these parameters, a body model can be determined. After the body model is determined, the system uses a smart wearable device to acquire the user's changing parameters in real time, including heart rate, exercise volume, and blood oxygen saturation. Exercise volume is determined by the instantaneous step count, which is directly proportional to the instantaneous step count. Based on these changing parameters, the system determines the display parameters for the body model and inserts these parameters into the body model to obtain the user's body profile.

[0081] In the technical solution of this invention, regarding the process of determining the body model, this application adopts a matching and reading method, eliminating the need for a modeling process. This method is very fast, but it requires staff to pre-build a body model library. First, the user determines some basic parameters, and then models these basic parameters to obtain a body model corresponding to each basic parameter. When it is necessary to model each user, the closest body model is queried from the body model library based on the user's basic parameters, and this model is used as the user's body model. The closest body model is the model with the smallest difference in basic parameters. In this architecture, the body model library contains basic parameter items and body model items. The more data items there are, the higher the accuracy, and the more accurate the matched body model.

[0082] The variable parameters include heart rate, exercise intensity, and blood oxygen saturation. Heart rate, exercise intensity, and blood oxygen saturation are all converted to the range of 0 to 255 and used as H value, S value, and V value, respectively, to obtain the display parameters. HSV space is a conventional color mode and is a commonly used color value space in computer display.

[0083] As a preferred embodiment of the technical solution of the present invention, the steps of acquiring user behavioral data with a time span, acquiring change information of body profile based on the time span, and training a mapping model from behavioral data to change information include:

[0084] Based on real-time acquisition of user change parameters by smart wearable devices, the change parameters are identified to determine the user's behavior type at each time point;

[0085] Within a time period, the behavior types at various moments are merged to obtain the behavior data for each time period;

[0086] Obtain information on changes in body profiles at each time period, use behavioral data from the same time period as features, and use information on changes in body profiles as labels to construct samples;

[0087] Collect a predetermined number of samples, construct a sample set, and divide the sample set into a training set and a test set according to a predetermined ratio to train the neural network model;

[0088] When the error rate of the neural network model is less than a preset error rate threshold, the neural network model is output as a mapping model from behavioral data to change information.

[0089] In one example of the technical solution of the present invention, the process of acquiring and applying change information is described. Based on the real-time acquisition of the user's changing parameters by the smart wearable device, the changing parameters are identified to determine the user's behavior type at each time. The correspondence between the changing parameters and the behavior type is known. In the prior art, there are already solutions for determining whether the user is sleeping, walking or sitting / lying down based on the changing parameters, which will not be described in detail in this application.

[0090] Within a time period, the behavior types at each moment are merged. A time period is generally a whole, that is, from 0:00 to 24:00. The process of merging the behavior types at each moment is as follows: the behavior types of adjacent moments are compared one by one. When the behavior types are the same, the two adjacent moments are merged. When the behavior types are different, the two adjacent moments are marked as abnormal moments. After the merging process of adjacent moments is executed, for abnormal moments, the closest merged behavior type of the abnormal moment is obtained, and the abnormal moment is merged into the closest merged behavior type. Only then is the merging process truly completed.

[0091] As a preferred embodiment of the technical solution of the present invention, the step of clustering users by comparing the latest body profiles and the latest mapping models of different users includes:

[0092] Obtain the body profile and mapping model of each user at the current moment;

[0093] Compare the body profiles and mapping models of different users, and calculate the profile distance and model distance;

[0094] Users are clustered based on the portrait distance and model distance.

[0095] In one example of the technical solution of this invention, the user clustering process is described. The body profile and mapping model of each user at the current moment are obtained. The current moment means the latest moment. The body profiles and mapping models of different users are compared, and the profile distance and model distance are calculated. The profile distance and model distance are used as parameters in the clustering process, and the clustering process is executed.

[0096] As an example of the technical solution of the present invention, the step of clustering users based on the profile distance and the model distance includes:

[0097] For any two users, read the body profiles of the two users, calibrate the origin of the two body profiles, and calculate the distance between the profiles;

[0098] Read the mapping models of two users, randomly determine a preset number of test data, input the test data into the two mapping models, obtain the output, and calculate the model distance based on the output;

[0099] User similarity is calculated based on image distance and model distance, and the users are clustered using the optics clustering algorithm; the similarity is inversely proportional to both image distance and model distance.

[0100] The calculation process for the image distance is as follows:

[0101] d1 represents the image distance; N, M, and K represent the dimensions of the space where the body model is located; A(i,j,k) and B(i,j,k) represent the parameters at point (i,j,k) in the two body models, respectively. A(i,j,k) and B(i,j,k) are both arrays used to represent the display parameters. When a body model does not display parameters at point (i,j,k), an array with all zero elements is used as a substitute array; dis{A(i,j,k),B(i,j,k)} represents the array distance between the two arrays; α is a preset adjustment coefficient.

[0102] The above content explains the clustering process. The clustering process adopts an unsupervised clustering scheme and does not limit the number of classes. When applying the optics clustering algorithm, the similarity needs to be determined first. The similarity is determined by the image distance and the model distance. It is worth mentioning that the degree of influence of the image distance and the model distance on the similarity can be determined by the preset weights.

[0103] Furthermore, regarding image distance, which is the difference between two body images, in practical applications, a cylindrical space is created starting from the origin, such as 2m*2m*3m. This space is sufficient to include most of the human body. The cylindrical space is scaled according to the scale of the body images. Then, positions are queried sequentially within the cylindrical space to determine whether the position belongs to a body image. If it does, the corresponding display parameters are determined. The display parameters of the two body images at that position are compared to obtain the distance at that position. Since the display parameters are arrays, the distance is calculated using array distance. After all array distances are calculated, the average of the array distances is calculated, and then a correction coefficient is introduced to obtain the final image distance.

[0104] Furthermore, regarding model distance, model distance is a test architecture where some test data are randomly selected, input into the mapping model, and after obtaining the output, they are compared, the difference is calculated, and the mean of the difference corresponding to each test data is calculated to obtain the final model distance. Since the output of the mapping model is the change information of the body portrait, its data structure is the same as that of the body portrait, so the formula for calculating portrait distance can also be applied.

[0105] As a preferred embodiment of the technical solution of the present invention, the step of generating a prompt message based on the physical examination data and broadcasting it to users of the same type when receiving physical examination data uploaded by any user includes:

[0106] Regularly send requests to various users to obtain physical examination data;

[0107] When any user uploads physical examination data, the data is transcoded to obtain digital physical examination data.

[0108] Locate abnormal data in digital health checkup data and insert a prompt box into the abnormal data;

[0109] Broadcast digital health check data containing prompts to similar users.

[0110] In the original technical solution, the physical examination data was entirely uploaded by the user. However, in reality, users may forget to upload their physical examination data after the examination. Therefore, this application periodically sends physical examination data retrieval requests to various users to remind them. When physical examination data uploaded by any user is received, the data is transcoded to obtain digital physical examination data. The purpose of this process is to standardize the format of physical examination data, because physical examination reports issued by different medical examination institutions may be different. Some are in tables, some are paper reports, and some are directly digital reports. This application standardizes all formats of physical examination data, and the standardized physical examination data is called digital physical examination data.

[0111] In digital health check data, abnormal data is located by comparing various indicators in the health check data with preset ranges. If the health check data exceeds the threshold, it is considered abnormal data. A prompt box is inserted into the abnormal data as a prompt message. Finally, the digital health check data containing the prompt box is broadcast to similar users.

[0112] Figure 2 A structural diagram of an artificial intelligence big data smart health management system is shown. In a preferred embodiment of the technical solution of the present invention, an artificial intelligence big data smart health management system is also provided, the system 10 comprising:

[0113] The body profile construction module 11 is used to acquire the user's body parameters based on smart devices and construct the user's body profile based on the body parameters.

[0114] The mapping model training module 12 is used to acquire user behavior data with a time span, obtain change information of body profile based on the time span, and train a mapping model from behavior data to change information.

[0115] User clustering module 13 is used to cluster users by comparing the latest physical profiles and the latest mapping models of different users at the latest time.

[0116] The physical examination data sharing module 14 is used to generate a prompt message based on the physical examination data when it receives physical examination data uploaded by any user of any type, and broadcast it to users of the same type.

[0117] Furthermore, the body portrait construction module 11 includes:

[0118] The basic parameter acquisition unit is used to acquire the user's basic parameters based on the smart body fat scale.

[0119] A body model matching unit is used to traverse and match body models in a preset body model library based on the basic parameters.

[0120] The variable parameter determination unit is used to acquire the user's variable parameters in real time based on the smart wearable device, and determine the display parameters of the body model based on the variable parameters.

[0121] The display parameter application unit is used to insert display parameters into the body model to obtain the user's body profile.

[0122] The body model library includes:; the traversal matching process is: different weights.

[0123] Specifically, the mapping model training module 12 includes:

[0124] The variable parameter identification unit is used to acquire the user's variable parameters in real time based on the smart wearable device, identify the variable parameters, and determine the user's behavior type at each time.

[0125] The behavior type merging unit is used to merge the behavior types at various times within a time period to obtain the behavior data for each time period.

[0126] The sample creation unit is used to obtain the change information of the body profile at each time period, and construct samples by using the behavioral data of the same time period as features and the change information of the body profile as labels.

[0127] The model training unit is used to count a preset number of samples, construct a sample set, and divide the sample set into a training set and a test set according to a preset ratio to train the neural network model.

[0128] The model output unit is used to output the neural network model as a mapping model from behavioral data to change information when the error rate of the neural network model is less than a preset error rate threshold.

[0129] Furthermore, the user clustering module 13 includes:

[0130] The parameter acquisition unit is used to acquire the body profile and mapping model of each user at the current moment;

[0131] The user comparison unit is used to compare the body profiles and mapping models of different users and calculate the profile distance and model distance.

[0132] The clustering execution unit is used to cluster users based on the profile distance and the model distance.

[0133] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. An artificial intelligence big data wisdom health management method, characterized in that, The method comprises: Obtaining the body parameters of the user based on the intelligent device, and constructing the body portrait of the user according to the body parameters; Obtaining the behavior data of the user containing a time span, obtaining the change information of the body portrait according to the time span, and training the mapping model of the behavior data to the change information; Comparing the latest body portrait and the latest mapping model of different users, and clustering the users; For any class of users, when receiving the physical examination data uploaded by any user, generating prompt information according to the physical examination data, and broadcasting to the same class of users; The step of obtaining the body parameters of the user based on the intelligent device, and constructing the body portrait of the user according to the body parameters comprises: Obtaining the basic parameters of the user based on the intelligent body fat scale; According to the basic parameters, traversing and matching the body model in the preset body model library; Real-time acquisition of the variable parameters of the user based on the intelligent wearable device, and determination of the display parameters of the body model according to the variable parameters; Inserting the display parameters into the body model to obtain the body portrait of the user; The variable parameters include heart rate, exercise amount and blood oxygen saturation, and the heart rate, exercise amount and blood oxygen saturation are converted to the range of 0 to 255 to be respectively taken as H value, S value and V value to obtain the display parameters; The step of comparing the latest body portrait and the latest mapping model of different users, and clustering the users comprises: Obtaining the body portrait and the mapping model of each user at the current time; Comparing the body portraits and the mapping models of different users, calculating the portrait distance and the model distance; Clustering the users according to the portrait distance and the model distance; The step of clustering the users according to the portrait distance and the model distance comprises: For any two users, reading the body portraits of the two users, calibrating the origins of the two body portraits, and calculating the portrait distance; Reading the mapping models of the two users, randomly determining a preset number of test data, inputting the test data into the two mapping models to obtain the output, and calculating the model distance according to the output; According to the portrait distance and the model distance, the similarity of the users is calculated, and the optics clustering algorithm is applied to cluster the users; the similarity is inversely proportional to the portrait distance and the model distance; The calculation process of the portrait distance is: ; denotes the image distance, , and K denote the dimensions of the space in which the body models are located; and denote the parameters at the midpoints of the two body models, and are arrays that represent the display parameters, and when a body model does not have a display parameter at the point , an array with all elements being zero is used as a substitute array; denotes the array distance between the two arrays; is a preset adjustment coefficient;​ The step of, for any class of users, when receiving the physical examination data uploaded by any user, generating prompt information according to the physical examination data, and broadcasting to the same class of users comprises: Sending physical examination data acquisition requests to each class of users at regular intervals; When receiving the physical examination data uploaded by any user, transcoding the physical examination data to obtain digital physical examination data; Positioning the abnormal data in the digital physical examination data, and inserting a prompt box in the abnormal data; Broadcasting the digital physical examination data containing the prompt box to the same class of users. 2.The artificial intelligence big data wisdom health management method according to claim 1, characterized in that, The step of obtaining the behavior data of the user containing a time span, obtaining the change information of the body portrait according to the time span, and training the mapping model of the behavior data to the change information comprises: Real-time acquisition of the variable parameters of the user based on the intelligent wearable device, identification of the variable parameters, and determination of the behavior type of the user at each time; In a time period, the behavior types at each time are combined to obtain the behavior data of each period; Obtain the change information of the body image of each period, take the behavior data of the same period as the feature, and construct the sample with the change information of the body image as the label; Statistically obtain a preset number of samples, construct a sample set, split the sample set into a training set and a test set according to a preset proportion, and train a neural network model; When the error rate of the neural network model is less than a preset error rate threshold, output the neural network model as a mapping model of behavior data to change information.

3. An artificial intelligence big data wisdom health management system, characterized in that, The system comprises: A body image construction module configured to obtain body parameters of a user based on an intelligent device, and construct a body image of the user according to the body parameters; A mapping model training module configured to obtain behavior data of the user containing a time span, obtain change information of the body image according to the time span, and train a mapping model of the behavior data to the change information; A user clustering module configured to compare the latest body image and the latest mapping model of different users, and cluster the users; A physical examination data sharing module configured to, for any type of user, generate prompt information according to physical examination data uploaded by any user, and broadcast to users of the same type; The body image construction module comprises: A basic parameter obtaining unit configured to obtain basic parameters of a user based on an intelligent body fat scale; A body model matching unit configured to traverse and match a body model in a preset body model library according to the basic parameters; A change parameter determining unit configured to obtain change parameters of the user in real time based on an intelligent wearable device, and determine display parameters of the body model according to the change parameters; A display parameter application unit configured to insert the display parameters into the body model to obtain a body image of the user; The change parameters include heart rate, exercise amount, and blood oxygen saturation, which are all converted to the range of 0 to 255 to obtain the display parameters as H value, S value, and V value respectively; The user clustering module comprises: A parameter obtaining unit configured to obtain the body image and the mapping model of each user at the current time; A user comparison unit configured to compare the body images and the mapping models of different users, and calculate image distance and model distance; A clustering execution unit configured to cluster the users according to the image distance and the model distance; The content of clustering the users according to the image distance and the model distance comprises: For any two users, read the body images of the two users, calibrate the origins of the two body images, and calculate the image distance; Read the mapping models of the two users, randomly determine a preset number of test data, input the test data into the two mapping models to obtain the output, and calculate the model distance according to the output; Calculate the similarity of the users according to the image distance and the model distance, and apply the optics clustering algorithm to cluster the users; the similarity is inversely proportional to the image distance and the model distance; The calculation process of the image distance comprises: ; denotes the image distance, , and K denote the dimensions of the space in which the body models are located; and denote the parameters at the midpoints of the two body models, and are arrays that represent the display parameters, and when a body model does not have a display parameter at the point , an array with all elements being zero is used as a substitute array; denotes the array distance between the two arrays; is a preset adjustment coefficient;​ For any type of user, when receiving physical examination data uploaded by any user, the content of generating prompt information according to the physical examination data and broadcasting to users of the same type comprises: Send a physical examination data obtaining request to each type of user at a regular time. When receiving any user uploaded physical examination data, the physical examination data is transcoded to obtain digital physical examination data; The digital physical examination data is positioned to locate abnormal data, and a prompt box is inserted in the abnormal data; The digital physical examination data containing the prompt box is broadcast to the same type of users. 4.The artificial intelligence big data wisdom health management system according to claim 3, characterized in that, The mapping model training module comprises: A variable parameter identification unit is configured to acquire a variable parameter of a user in real time based on an intelligent wearable device, identify the variable parameter, and determine a behavior type of the user at each time; A behavior type merging unit is configured to merge the behavior types at each time in a time period to obtain behavior data of each period; A sample creation unit is configured to acquire change information of a body portrait of each period, use the behavior data of the same period as a feature, use the change information of the body portrait as a label to construct a sample, and use the sample to train a neural network model; A model output unit is configured to output the neural network model as a mapping model of behavior data to change information when an error rate of the neural network model is less than a preset error rate threshold. ​

Images