Health status assessment method, system and electronic device

Abstract

The present invention provides a health status assessment method, system and electronic device, which relates to the field of health status assessment. This method can collect and process heterogeneous data from multiple sources, and conduct in-depth fusion analysis on it to obtain the physique identification data of the user, so as to obtain the physique situation of the user and conduct health status analysis and evaluation on it to obtain accurate health assessment results, thus realizing the complete processing flow from data collection to health status assessment.

Description

Technical Field

[0001] The present invention relates to the field of health status assessment, and particularly to a health status assessment method, system and electronic device. Background Art

[0002] In the process of assessing the health status of users, traditional assessment methods use methods such as questionnaire processing, four diagnostic data acquisition, and daily monitoring data collection to collect data from users. Since it involves different fields, the data collection and processing process is complex, and it is difficult to uniformly process the health status assessment process of users. In addition, for the assessment process of users' health status, the existing technology mainly realizes it based on the semantic analysis results, lacking the fusion analysis of health data, resulting in poor accuracy of the health status assessment of users. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a health status assessment method, system and electronic device. This method can collect and process heterogeneous data from multiple sources, and conduct in-depth fusion analysis on it to obtain the physical constitution identification data of the user to be evaluated, so as to obtain the physical constitution of the user and conduct a health status analysis and evaluation on it to obtain an accurate health assessment result, realizing the complete processing flow from data collection to health status assessment, thus solving the above problems existing in the prior art.

[0004] In the first aspect, an embodiment of the present invention provides a health status assessment method, which includes:

[0005] Health data collection step: Collect the physical constitution monitoring data of the user to be evaluated based on a preset multi-source heterogeneous database, and determine the health data of the user to be evaluated based on the semantic result of the physical constitution monitoring data;

[0006] Physical constitution identification processing step: Obtain the feature vector corresponding to the health data, determine the weight matrix and bias vector corresponding to the feature vector, and use the feature vector, weight matrix and bias vector to determine the physical constitution identification data corresponding to the user to be evaluated;

[0007] Health data analysis step: Conduct density clustering calculation on the physical constitution identification data to obtain the health index clustering result of the user to be evaluated, and calculate the health index score of the user to be evaluated using the health index clustering result;

[0008] Assessment result generation step: Determine the health risk level of the user to be evaluated according to the health index score, and generate the health assessment result of the user to be evaluated based on the judgment matrix corresponding to the health risk level.

[0009] Optionally, the health data collection step includes:

[0010] Determine the constitution questionnaire database, four diagnostic databases, and health monitoring database based on multi-source heterogeneous databases;

[0011] Obtain the constitution monitoring data corresponding to the user to be evaluated by using the constitution questionnaire database, four diagnostic databases, and health monitoring database;

[0012] Construct a random forest corresponding to the constitution monitoring data, calculate the anomaly score of each data point in the random forest, and after updating and processing the constitution monitoring data according to the anomaly score, obtain the semantic relationship corresponding to the constitution monitoring data;

[0013] Determine the semantic result corresponding to the constitution monitoring data by using the semantic relationship, and determine the health data of the user to be evaluated according to the semantic result.

[0014] Optionally, obtaining the constitution monitoring data corresponding to the user to be evaluated by using the constitution questionnaire database, four diagnostic databases, and health monitoring database includes:

[0015] Obtain the constitution questionnaire data corresponding to the user to be evaluated by using the constitution questionnaire database, and determine the first constitution monitoring data corresponding to the user to be evaluated according to the weight value and semantic result corresponding to the constitution questionnaire data;

[0016] Obtain the pulse condition data, tongue diagnosis data, voice data, and body temperature data corresponding to the user to be evaluated by using the four diagnostic databases, and determine the second constitution monitoring data corresponding to the user to be evaluated according to the feature vectors corresponding to the pulse condition data, tongue diagnosis data, voice data, and body temperature data;

[0017] Obtain the health monitoring data and environmental data corresponding to the user to be evaluated by using the health monitoring database, and determine the third constitution monitoring data corresponding to the user to be evaluated according to the health monitoring data and environmental data;

[0018] Determine the constitution monitoring data according to the first constitution monitoring data, the second constitution monitoring data, and the third constitution monitoring data.

[0019] Optionally, the constitution identification processing steps include:

[0020] Obtain the text data corresponding to the health data, generate the label sequence data corresponding to the text data based on the initialized bidirectional long short-term memory model and conditional random field model, and determine the feature vector corresponding to the health data by using the label sequence data;

[0021] Input the feature vector into the trained deep belief network, control the deep belief network to obtain the modal data corresponding to the feature vector, and determine the weight matrix and bias vector corresponding to the feature vector by using the joint representation result between the modal data;

[0022] Determine the hidden units corresponding to the deep belief network according to the preset activation function, using the feature vector, weight matrix, and bias vector, and use the hidden units to output the physique identification data corresponding to the user to be evaluated under the health data.

[0023] Optionally, the health data analysis step includes:

[0024] Obtain the physique type data and disease type data corresponding to the user to be evaluated based on the physique identification data, and obtain the association rules corresponding to the physique type data and disease type data;

[0025] Use the association rules to perform density clustering calculation on the physique identification data, obtain the characteristic distribution result between the physique type data and disease type data of the user to be evaluated, and determine the health index clustering result of the user to be evaluated based on the characteristic distribution result;

[0026] Construct a prediction model corresponding to the physique type data and disease type data according to the initialized Bayesian network, and use the prediction model to calculate the health index score corresponding to the health index clustering result.

[0027] Optionally, constructing a prediction model corresponding to the physique type data and disease type data according to the initialized Bayesian network includes:

[0028] Obtain the first node set corresponding to the physique type data and the second node set corresponding to the disease type data;

[0029] Construct the corresponding conditional probability table of the Bayesian network based on the first node set and the second node set, and use the conditional probability table to determine the structure parameters of the Bayesian network;

[0030] Construct a prediction model based on the structure parameters and the Bayesian network.

[0031] Optionally, the evaluation result generation step includes:

[0032] Obtain the window function corresponding to the health index clustering result, calculate the mean result of the health index score under the window function, and determine the health risk level of the user to be evaluated according to the mean result;

[0033] Construct an evaluation index strategy corresponding to the user to be evaluated based on the health index clustering result, and use the evaluation index strategy to determine the judgment matrix;

[0034] Determine the health evaluation result corresponding to the user to be evaluated according to the judgment matrix.

[0035] Optionally, determining the health evaluation result corresponding to the user to be evaluated according to the judgment matrix includes:

[0036] Calculate the maximum eigenvalue of the judgment matrix and its corresponding eigenvector;

[0037] Obtain the index weight vector of the judgment matrix by using the maximum eigenvalue and the corresponding eigenvector;

[0038] Determine the membership degree vector corresponding to the index weight vector according to the health risk level, and construct the fuzzy relation matrix corresponding to the judgment matrix based on the membership degree vector;

[0039] Generate the evaluation vector corresponding to the user to be evaluated according to the fuzzy relation matrix, and determine the health evaluation result by using the evaluation vector.

[0040] In a second aspect, the present invention provides a health status evaluation system, which includes:

[0041] A health data collection unit, configured to collect the physical fitness monitoring data of the user to be evaluated based on a preset multi-source heterogeneous database, and determine the health data of the user to be evaluated based on the semantic result of the physical fitness monitoring data;

[0042] A physical constitution identification processing unit, configured to obtain the feature vector corresponding to the health data, determine the weight matrix and bias vector corresponding to the feature vector, and determine the physical constitution identification data corresponding to the user to be evaluated by using the feature vector, weight matrix and bias vector;

[0043] A health data analysis unit, configured to perform density clustering calculation on the physical constitution identification data to obtain the health index clustering result of the user to be evaluated, and calculate the health index score of the user to be evaluated by using the health index clustering result;

[0044] An evaluation result generation unit, configured to determine the health risk level of the user to be evaluated according to the health index score, and generate the health evaluation result of the user to be evaluated based on the judgment matrix corresponding to the health risk level.

[0045] In a third aspect, an embodiment of the present invention further provides an electronic device, including a processor and a memory, where the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the steps of the health status evaluation method provided in the first aspect.

[0046] In a fourth aspect, an embodiment of the present invention further provides a storage medium, which stores computer executable instructions. When the computer executable instructions are called and executed by a processor, the computer executable instructions cause the processor to implement the steps of the health status evaluation method provided in the first aspect.

[0047] A health status assessment method, system and electronic device provided by an embodiment of the present invention, in the process of assessing the health status of a user, first collects physical fitness monitoring data of the user to be evaluated based on a preset multi-source heterogeneous database, and determines the health data of the user to be evaluated based on the semantic results of the physical fitness monitoring data, so as to complete the health data collection process; then obtains the feature vector corresponding to the health data, determines the weight matrix and bias vector corresponding to the feature vector, and uses the feature vector, weight matrix and bias vector to determine the physical fitness identification data corresponding to the user to be evaluated, so as to complete the physical fitness identification process; then performs density clustering calculation on the physical fitness identification data to obtain the health index clustering result of the user to be evaluated, and calculates the health index score of the user to be evaluated by using the health index clustering result. Thus, the health data analysis process is completed; finally, the health risk level of the user to be evaluated is determined according to the health index score, and the health assessment result of the user to be evaluated is generated based on the judgment matrix corresponding to the health risk level. This method can collect and process heterogeneous data from multiple sources, and perform in-depth fusion analysis on them to obtain the physical fitness identification data of the user, so as to obtain the physical fitness of the user and perform health status analysis and evaluation on it to obtain an accurate health assessment result, thus realizing the complete processing flow from data collection to health status assessment.

[0048] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention are achieved and obtained by the structures specifically pointed out in the specification, claims and drawings.

[0049] To make the above objectives, features and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given, and in conjunction with the accompanying drawings, the detailed description is as follows. Brief Description of the Drawings

[0050] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0051] Figure 1 It is a flowchart of a health status assessment method provided by an embodiment of the present invention;

[0052] Figure 2 It is a flowchart of step S101 of health data collection in a health status assessment method provided by an embodiment of the present invention;

[0053] Figure 3It is a flowchart of step S202 in a health status assessment method provided by an embodiment of the present invention;

[0054] Figure 4 It is a flowchart of the physical constitution identification processing step S102 in a health status assessment method provided by an embodiment of the present invention;

[0055] Figure 5 It is a flowchart of the health data analysis step S103 in a health status assessment method provided by an embodiment of the present invention;

[0056] Figure 6 It is a flowchart of constructing a prediction model corresponding to physical constitution type data and disease type data according to an initialized Bayesian network in a health status assessment method provided by an embodiment of the present invention;

[0057] Figure 7 It is a flowchart of the assessment result generation step S104 in a health status assessment method provided by an embodiment of the present invention;

[0058] Figure 8 It is a flowchart of step S703 in a health status assessment method provided by an embodiment of the present invention;

[0059] Figure 9 It is a schematic diagram of a health status assessment system provided by an embodiment of the present invention;

[0060] Figure 10 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention.

[0061] Icon:

[0062] 910 - Health data acquisition unit; 920 - Physical constitution identification processing unit; 930 - Health data analysis unit; 940 - Assessment result generation unit;

[0063] 101 - Processor; 102 - Memory; 103 - Bus; 104 - Communication interface. Detailed implementation manners

[0064] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0065] In the process of evaluating the health status of users, traditional evaluation methods use methods such as questionnaire processing, four diagnostic data acquisition, and daily monitoring data collection to collect data from users. Since it involves different fields, the data collection and processing process is complex, making it difficult to uniformly process the health status evaluation process of users. In addition, for the evaluation process of users' health status, the existing technologies mainly rely on semantic analysis results to achieve, lacking the fusion analysis of health data, resulting in poor accuracy of the evaluation of users' health status. Based on this, the embodiments of the present invention provide a health status evaluation method, system, and electronic device. This method can collect and process heterogeneous data from multiple sources, perform in-depth fusion analysis on them to obtain the physique identification data of the user, thereby obtaining the physique of the user and conducting a health status analysis and evaluation on it to obtain accurate health evaluation results, thus realizing the complete processing flow from data collection to health status evaluation.

[0066] To facilitate the understanding of this embodiment, first, a health status evaluation method disclosed in the embodiments of the present invention will be introduced in detail. This method is as Figure 1 shown and includes:

[0067] Health data collection step S101: Collect the physique monitoring data of the user to be evaluated based on a preset multi-source heterogeneous database, and determine the health data of the user to be evaluated based on the semantic results of the physique monitoring data;

[0068] Physique identification processing step S102: Obtain the feature vector corresponding to the health data, determine the weight matrix and bias vector corresponding to the feature vector, and use the feature vector, weight matrix, and bias vector to determine the physique identification data corresponding to the user to be evaluated;

[0069] Health data analysis step S103: Perform density clustering calculation on the physique identification data to obtain the health index clustering result of the user to be evaluated, and calculate the health index score of the user to be evaluated using the health index clustering result;

[0070] Evaluation result generation step S104: Determine the health risk level of the user to be evaluated according to the health index score, and generate the health evaluation result of the user to be evaluated based on the judgment matrix corresponding to the health risk level.

[0071] The main purpose of the health data collection step is to realize the collection process of the health data of the user to be evaluated. It mainly relies on a multi-source heterogeneous database to collect the physique monitoring data of the user. In the actual scenario, the collected physique monitoring data can be cleaned and integrated, and the health data of the user to be evaluated can be obtained based on the semantic results of the physique monitoring data.

[0072] Subsequently, a physical constitution identification processing step is performed. The main purpose of this step is to identify the user's physical constitution. After vectorizing the health data, the corresponding feature vector is obtained, and then the weight matrix and bias vector corresponding to the feature vector are obtained. Using the above vectors and matrices, a relevant physical constitution identification model is constructed, and the physical constitution identification data of the user to be evaluated is obtained using this model.

[0073] The main purpose of the health data analysis step is to analyze the physical constitution identification data to obtain the specific health condition of the user to be evaluated, which is specifically characterized by the health index score. Specifically, by performing density clustering calculation on the physical constitution identification data, the health index score of the user in the current physical constitution state is obtained using the obtained health index clustering result, realizing the process of quantifying the health index represented by the physical constitution identification data of the user.

[0074] Finally, an evaluation result generation step is performed. This step mainly classifies the levels according to the quantitative score result of the health index, thereby obtaining the health risk level of the user to be evaluated. Then, based on the corresponding judgment matrix for different health risk levels, the health evaluation result of the user is generated. It is worth mentioning that the judgment matrix represents the parameters of the importance of relevant health indicators in the health risk level. Using the relevant eigenvectors and eigenvalue results of the judgment matrix for fuzzy operation, the health evaluation result of the user to be evaluated is obtained.

[0075] Optionally, the health data collection step S101, as Figure 2 shown, includes:

[0076] Step S201, determining the physical constitution questionnaire database, the four diagnostic databases, and the health monitoring database based on the multi-source heterogeneous database;

[0077] Step S202, obtaining the physical constitution monitoring data corresponding to the user to be evaluated using the physical constitution questionnaire database, the four diagnostic databases, and the health monitoring database;

[0078] Step S203, constructing a random forest corresponding to the physical constitution monitoring data, calculating the anomaly score of each data point in the random forest, and after updating the physical constitution monitoring data according to the anomaly score, obtaining the semantic relationship corresponding to the physical constitution monitoring data;

[0079] Step S204, determining the semantic result corresponding to the physical constitution monitoring data using the semantic relationship, and determining the health data of the user to be evaluated according to the semantic result.

[0080] Specifically, the health data collection step mines the constitution questionnaire data, four diagnostic data, and health monitoring data, which are respectively implemented based on the corresponding constitution questionnaire database, four diagnostic database, and health monitoring database in the multi-source heterogeneous database. After the constitution questionnaire database, four diagnostic database, and health monitoring database are determined, the corresponding constitution questionnaire data, four diagnostic data, and health monitoring data can be obtained through the corresponding database. Then, after collecting, optimizing, and mining these data, the constitution monitoring data of the user to be evaluated is obtained.

[0081] After obtaining the constitution monitoring data, data cleaning processing is performed on it. After constructing a random forest corresponding to the constitution monitoring data, the outliers in the constitution monitoring data are determined by calculating the anomaly score of each data point in the random forest. Specifically, by constructing a random forest to calculate the anomaly score s(a,b) of each data point, the formula used is , where E(h(a)) is the average path length of data point a in the random forest, and c(b) is the normalization factor, which is determined according to the data volume b. The higher the score, the more abnormal the data point, so as to accurately identify and remove the error data.

[0082] In the actual scenario, a data consistency check algorithm can be used to verify the consistency of data from different sources. For example, compare the height and weight data filled in the questionnaire with the data collected by devices such as body fat scales. If the difference exceeds a certain threshold, secondary verification is carried out or marked as suspicious data for further manual review.

[0083] After updating the constitution monitoring data, map the constitution questionnaire data, four diagnostic data, and health monitoring data to a unified semantic framework to obtain the semantic relationship between these data. For example, semantic associations such as "has symptoms" and "corresponding pulse condition" can be defined between "yang deficiency constitution" and data such as "fear of cold symptoms" and "deep and thready pulse". Through semantic reasoning, deep integration and correlation analysis of the data are realized, and the semantic results contained therein are obtained based on the semantic relationship, so as to obtain the health data of the user to be evaluated.

[0084] After obtaining the health data, its security can be improved by encrypting it and storing it in a specific storage architecture. Specifically, a hybrid encryption algorithm can be used. For example, first use an asymmetric encryption algorithm (such as RSA) to encrypt and transmit the AES encryption key to ensure the security of the key; then use the AES algorithm to efficiently encrypt and store the data. When decrypting the data, first use the private key to decrypt the AES key, and then use the AES key to decrypt the data to improve the security and efficiency of the encryption system. In the actual scenario, homomorphic encryption technology can also be introduced to allow specific operations on encrypted data. For example, perform statistical analysis on the encrypted health data (such as calculating the average blood pressure of a certain constitution type population) without decrypting the data first, further protecting the data privacy.

[0085] For the storage architecture, a distributed storage architecture can be constructed to disperse the encrypted data and store it on multiple nodes. For example, a distributed ledger using blockchain technology stores some key data digests, leveraging its immutability feature to ensure data integrity and traceability. At the same time, different copies are stored on the cloud and local servers respectively, improving data reliability and availability through redundant storage to prevent data loss. In practical scenarios, a strict access control policy model can also be established, combining role-based access control (RBAC) with attribute-based encryption (ABE). For example, different access permission attributes are set for different roles such as medical staff, health managers, and users. Only users who meet specific attribute conditions (such as doctor role and specific department qualifications) can decrypt and access the corresponding data, further refining access permission management to ensure data security.

[0086] Therefore, in the health data collection step, it is crucial to obtain the physical fitness monitoring data corresponding to the user to be evaluated using the physical fitness questionnaire database, the four diagnostic databases, and the health monitoring database. Optionally, the step S202 of obtaining the physical fitness monitoring data corresponding to the user to be evaluated using the physical fitness questionnaire database, the four diagnostic databases, and the health monitoring database is as Figure 3 shown, including:

[0087] Step S301: Obtain the physical fitness questionnaire data corresponding to the user to be evaluated using the physical fitness questionnaire database, and determine the first physical fitness monitoring data corresponding to the user to be evaluated according to the weight value and semantic result corresponding to the physical fitness questionnaire data;

[0088] Step S302: Obtain the pulse condition data, tongue diagnosis data, voice data, and body temperature data corresponding to the user to be evaluated using the four diagnostic databases, and determine the second physical fitness monitoring data corresponding to the user to be evaluated according to the feature vectors corresponding to the pulse condition data, tongue diagnosis data, voice data, and body temperature data;

[0089] Step S303: Obtain the health monitoring data and environmental data corresponding to the user to be evaluated using the health monitoring database, and determine the third physical fitness monitoring data corresponding to the user to be evaluated according to the health monitoring data and environmental data;

[0090] Step S304: Determine the physical fitness monitoring data according to the first physical fitness monitoring data, the second physical fitness monitoring data, and the third physical fitness monitoring data.

[0091] For the physical fitness questionnaire, after obtaining the physical fitness questionnaire data corresponding to the user to be evaluated using the physical fitness questionnaire database, the weight coefficient method can be used to assign different weights to each question according to its importance for physical fitness judgment. For example, a higher weight of 0.8 is assigned to the key question "Do you have cold hands and feet all year round" for judging yang deficiency constitution, while the weight of the general question "Do you occasionally have headaches" is set to 0.2. Calculate the questionnaire score through the weighted summation formula: , where Q is the total score of the questionnaire, is the weight of the i-th question, is the answer score of the i-th question (e.g., 0 means no, 1 means yes); and then the physical constitution tendency is preliminarily divided according to the total score range. In the actual processing process, fuzzy logic algorithms can be used to handle some fuzzy problems. For example, for the question of "degree of dietary preference", set fuzzy options such as "very like", "like", "general", "dislike", "very dislike", etc., and convert these options into precise numerical values through fuzzy membership functions for better quantitative data analysis.

[0092] For the four diagnostic data, the corresponding pulse data and tongue diagnosis data can be obtained through a pulse meter and a tongue diagnosis instrument. On this basis, voice data and temperature data are also considered. Specifically, by adding an interface for the auscultation device to collect the user's voice (such as whether the voice is low, whether there is coughing and wheezing, etc.) and odor (such as breath, body odor, etc.) data, and then extracting the features of these data, such as the frequency and amplitude of the voice, chemical composition analysis of the odor, etc., and finally converting them into digital feature vectors. For example, the voice frequency feature f can be calculated by the fast Fourier transform (FFT) formula where f(n) is the discrete sampling value of the voice signal, N is the number of sampling points, k is the frequency index, and j is the imaginary number. These feature vectors are stored together with other four diagnostic data to provide richer information for physical constitution identification.

[0093] For the health monitoring data, time series analysis algorithms can be used to deeply mine the daily health monitoring data. For example, analyze the sleep data through the autoregressive moving average (ARMA) model , where S t is the sleep data at the current moment (such as sleep time, sleep depth, etc.), S t-i is the sleep data at the past i-th moment, and are the model parameters, is the white noise error term; the future sleep trend of the user is predicted through this model to discover potential sleep problems in advance. In addition, by obtaining environmental data, such as local air quality, temperature, humidity and other information, it is associated with the user's health monitoring data. For example, it is found that during haze weather, the respiratory symptoms of certain physical constitution groups (such as phlegm-dampness constitution) will worsen, and more comprehensive basis for health management is provided through correlation analysis.

[0094] Optionally, the physical constitution identification processing step S102, as Figure 4 shown, includes:

[0095] Step S401, obtain the text data corresponding to the health data, generate the label sequence data corresponding to the text data based on the initialized bidirectional long short-term memory model and conditional random field model, and use the label sequence data to determine the feature vector corresponding to the health data;

[0096] Step S402, input the feature vector into the trained deep belief network, control the deep belief network to obtain the modal data corresponding to the feature vector, and use the joint representation result between the modal data to determine the weight matrix and bias vector corresponding to the feature vector;

[0097] Step S403, according to the preset activation function and using the feature vector, weight matrix and bias vector, determine the hidden units corresponding to the deep belief network, and use the hidden units to output the physical constitution identification data corresponding to the user to be evaluated under the health data.

[0098] In the process of obtaining the physical constitution identification data, corresponding questionnaires can be set based on the theory of physical constitution for assistance. Specifically, the analytic hierarchy process (AHP) can be used to determine the weight distribution of the questionnaire questions. First, construct the judgment matrix A, where the element α in matrix A ij represents the degree of importance of question i relative to question j for physical constitution judgment. By calculating the maximum eigenvalue of matrix A and its corresponding eigenvector, the relative weight w of each question is obtained i . This can more scientifically measure the contribution of each question in physical constitution identification and make the analysis of questionnaire results more accurate. In actual scenarios, context-based question design can be introduced to simulate different life scenarios and ask users about their physical reactions. For example, "In the cold winter, how is the warmth of your hands and feet?" Such context-based questions can more intuitively obtain information related to physical constitution and reduce user understanding deviation.

[0099] After obtaining the text data corresponding to the health data using the questionnaire, perform natural language processing on it. Specifically, use the initialized bidirectional long short-term memory model and conditional random field model to obtain the label sequence data corresponding to the text data, and use the label sequence data to determine the feature vector corresponding to the health data. Specifically, the bidirectional long short-term memory network (BiLSTM) can be combined with the conditional random field (CRF) for key information extraction. BiLSTM can effectively capture the context information of the text sequence, and CRF is used to optimize the output of BiLSTM to obtain a more accurate label sequence (i.e., key information). Let the input text sequence be X=(x1,x2,……,x n ), the hidden layer output of BiLSTM be H=(h1,h2,……,h n ), and the transition matrix of CRF be T. Then the probability calculation formula for the label sequence Y=(y1,y2,……,y n ) is: , where is the scoring function output by the BiLSTM, is the transition scoring function of the CRF, Y x is the set of all possible tag sequences, is a specific tag sequence in Y x in is to the set Y x all possible tag sequences in perform a summation operation.

[0100] Pre-train the natural language processing model by building a corpus of physical constitution corresponding to the relevant medical system. The corpus can include text data such as traditional Chinese medicine ancient books, clinical cases, and physical constitution research literature. After annotation, it is used for model training to enable the model to better understand professional terms and semantic relationships, and improve the accuracy and efficiency of key information extraction. The generated model can be a physical constitution identification model constructed by the ensemble learning method. For example, integrate multiple machine learning algorithms such as support vector machine (SVM), decision tree (DT), and random forest (RF). Let M1, M2, ……, M k be k different basic models. For the input user data x, the prediction result of each model is y i = M i (x). The final physical constitution identification result is obtained through the weighted voting method, where δ(y i , j) is the indicator function, which is 1 when y i = j, and 0 otherwise. w m i is the weight of the model M i and is determined by methods such as cross-validation.

[0101] After obtaining the feature vectors, input them into the deep belief network (DBN) for data fusion. In a specific scenario, the physical constitution questionnaire data, traditional Chinese medicine four diagnosis data, daily health monitoring data, and living habit data are respectively input into the DBN as different modalities. Let v q , v s , v m , v h be the feature vectors of the questionnaire, four diagnosis, health monitoring, and living habit data respectively. The DBN learns the joint representation between different modality data through layer-by-layer unsupervised pre-training and supervised fine-tuning, and finally outputs the physical constitution identification result. During the pre-training process, the hidden units of each layer are calculated by the following formula: , where W l is the weight matrix, b l is the bias vector, is the activation function.

[0102] Subsequently, a combination of feature-level fusion and decision-level fusion is adopted. In feature-level fusion, the features of different source data are normalized and then concatenated into a joint feature vector, which is then input into the physical constitution identification model. In decision-level fusion, independent physical constitution identification sub-models are first trained using different source data, and then the output results of the sub-models are fused through methods such as fuzzy integral to obtain the final physical constitution identification data.

[0103] Optionally, the health data analysis step S103, as Figure 5 shown, includes:

[0104] Step S501, obtaining the physical constitution type data and disease type data corresponding to the user to be evaluated based on the physical constitution identification data, and obtaining the association rules corresponding to the physical constitution type data and disease type data;

[0105] Step S502, using the association rules to perform density clustering calculation on the physical constitution identification data, obtaining the feature distribution result between the physical constitution type data and disease type data of the user to be evaluated, and determining the health index clustering result of the user to be evaluated based on the feature distribution result;

[0106] Step S503, constructing a prediction model corresponding to the physical constitution type data and disease type data according to the initialized Bayesian network, and calculating the health index score corresponding to the health index clustering result using the prediction model.

[0107] In the health data analysis step, an improved Apriori algorithm can be used for association rule mining to find strong association rules between physical constitution types and diseases. Based on the traditional Apriori algorithm, an interest degree index is introduced. Let U be the item set of physical constitution types, V be the item set of diseases, the support degree be S(U, V), the confidence degree be C(U, V), and the interest degree I(U, V)=S(U, V)×(C(U, V)-P(V)), where P(V) is the prior probability of disease V. By setting the minimum interest degree threshold, more valuable physical constitution-disease association rules can be mined, avoiding the mining of a large number of meaningless associations.

[0108] Then, the density clustering algorithm (such as DBSCAN) is used to perform clustering analysis on the physical constitution-disease data. Let the data point set Z={z1, z2, ……, z n}, for the data point z i , calculate the number of data points within its epsilon-neighborhood. If the number exceeds the set minimum number of points MinPts, then z i is a core point, and clusters are formed by continuously expanding the neighborhood of the core point. Through clustering analysis, the distribution characteristics of different physical constitution-disease groups are found, providing more detailed classification information for the prediction model.

[0109] Optionally, construct a prediction model corresponding to the physical constitution type data and the disease type data based on the initialized Bayesian network, such as Figure 6 shown, including:

[0110] Step S601, obtain the first node set corresponding to the physical constitution type data and the second node set corresponding to the disease type data;

[0111] Step S602, construct the corresponding conditional probability table of the Bayesian network based on the first node set and the second node set, and use the conditional probability table to determine the structure parameters of the Bayesian network;

[0112] Step S603, construct a prediction model based on the structure parameters and the Bayesian network.

[0113] The construction of the prediction model is based on the Bayesian network to construct a constitution-disease prediction model. Let the physical constitution type node set be T = {t1, t2, ……, t m}, the disease node set be D = {d1, d2, ……, d n}, and the other influencing factor node set be F = {f1, f2, ……, f k}. Determine the structure and parameters of the Bayesian network by learning a large amount of data, that is, the conditional probability table P(d i |t j , f l ), that is, the probability of the occurrence of disease d j under the condition that the physical constitution type is t l and the other influencing factor is f i . P(d i , t j , f l ): The joint probability of the simultaneous occurrence of disease d i , physical constitution type t j , and other influencing factor f l . P(t j , f l |d i ): The conditional probability of the physical constitution type being t i and the other influencing factor being f j under the condition that disease d l occurs. P(d i ): The prior probability of the occurrence of disease d i , that is, the probability of the occurrence of disease d i itself without considering other conditions. d i : The i-th disease in the disease node set D = {d1, d2, ……, d n}. t j : The physical constitution type node set T = {t1, t2, ……, t mthe j-th physical constitution type in {...}. f l : The set of other influencing factor nodes F = {f1, f2, ……, f k} the l-th other influencing factor in {...}. : The summation symbol, indicating the summation of i from 1 to n, that is, the summation of the relevant probabilities for all possible diseases d i to play a role in normalization, making the finally calculated conditional probability between 0 and 1 by summing up the relevant probabilities.

[0114] For the given user's physical constitution type and other influencing factor data, the probability of disease occurrence is calculated using Bayes' formula as follows:

[0115] .

[0116] To evaluate the performance of the prediction model using multiple indicators, in addition to the commonly used accuracy, recall rate, and F1 value, the Brier score is introduced. Let n be the number of prediction samples, p i be the probability of the i-th sample having a disease, and y i be whether the disease actually occurs (1 means it occurs, 0 means it does not occur). The Brier score formula is: . Specifically, n: the number of prediction samples, that is, the number of samples participating in the model prediction evaluation. p i The probability of the i-th sample having a disease predicted by the model, with a value range between 0 and 1. y i The identifier of whether the i-th sample actually has a disease, 1 means the sample actually has the disease, 0 means it does not have the disease. BS: The Brier score, used to measure the difference between the model's predicted probability and the actual result. The lower the Brier score, the better the model's prediction effect. By comprehensively evaluating the indicators, the model parameters and structure are continuously optimized to improve the prediction accuracy.

[0117] Optionally, the evaluation result generation step S104, as Figure 7 shown, includes:

[0118] Step S701, obtain the window function corresponding to the health index clustering result, calculate the mean result of the health index scores under the window function, and determine the health risk level of the user to be evaluated according to the mean result;

[0119] Step S702, construct the evaluation index strategy corresponding to the user to be evaluated based on the health index clustering result, and use the evaluation index strategy to determine the judgment matrix;

[0120] Step S703, determine the health evaluation result corresponding to the user to be evaluated according to the judgment matrix.

[0121] For the monitoring process of health index scores, the quantile regression method can be used to set the dynamic normal range of health indicators. Let the health indicator data be K, and the vector of influencing factors (such as constitution type, age, gender, etc.) be J. The quantile regression model is: , where q is the quantile level (such as 0.25, 0.5, 0.75, etc.), and β j,q is the corresponding regression coefficient. By establishing quantile regression models for different groups of people with different constitution types, ages, and genders respectively, the normal range boundaries of health indicators at different quantiles are determined to adapt to individual differences and dynamic changes.

[0122] 2) Establish an adaptive threshold adjustment mechanism to dynamically adjust the normal range threshold according to the changing trend of the user's health data. For example, using the exponential smoothing method, let the current threshold be T t , the new data point be x t , and the smoothing coefficient be α. Then the updated threshold is: T t+1 = αx t + (1 - α)T t . When the user's health indicator data shows a continuous upward or downward trend, the threshold can be adjusted accordingly to improve the sensitivity of abnormal fluctuation monitoring.

[0123] In the actual scenario, a streaming data processing framework such as Apache Flink can be used to efficiently process health data. In Flink, window functions are used to group and analyze the data. For example, using a sliding time window, let the window size be W and the sliding step be S. Then, statistical analysis (such as calculating the mean, standard deviation, etc.) is performed on the health indicator data within each time window. The formula used is: , where represents the sample mean of the health indicator data within a sliding time window; x i is the data point within the window, n is the number of data points within the window; s is the sample standard deviation of the health indicator data within the sliding time window. When the statistical indicator exceeds the dynamic normal range, an abnormal warning is triggered.

[0124] Specifically, a comprehensive evaluation model based on a fuzzy inference system can also be constructed. The user's constitution type, the degree of health indicator abnormality, symptom information, etc. are used as the inputs of the fuzzy inference system, and the output is the health risk level. For example, define the fuzzy sets of constitution type as (very consistent, relatively consistent, general, less consistent, inconsistent), the fuzzy sets of the degree of health indicator abnormality as (mild, moderate, severe), and the fuzzy sets of symptom information as (none, occasional, frequent). Through establishing a fuzzy rule table (such as "if the constitution type is phlegm-dampness constitution and the blood pressure abnormality degree is moderate and there are frequent dizziness symptoms, then the health risk level is high") for fuzzy inference, a more comprehensive and accurate health risk assessment result is obtained, providing strong support for early warning decision-making.

[0125] Optionally, step S703 of determining the health assessment result corresponding to the user to be evaluated according to the judgment matrix is as follows Figure 8 shown in the figure and includes

[0126] Step S801: Calculate the maximum eigenvalue of the judgment matrix and its corresponding eigenvector;

[0127] Step S802: Obtain the index weight vector of the judgment matrix by using the maximum eigenvalue and its corresponding eigenvector;

[0128] Step S803: Determine the membership degree vector corresponding to the index weight vector according to the health risk level, and construct the fuzzy relation matrix corresponding to the judgment matrix according to the membership degree vector;

[0129] Step S804: Generate the evaluation vector corresponding to the user to be evaluated according to the fuzzy relation matrix, and determine the health assessment result by using the evaluation vector.

[0130] The generation of the health assessment result can be realized based on a relevant assessment model, which is an assessment model combining the analytic hierarchy process (AHP) and the fuzzy comprehensive evaluation method, such as physical constitution types (yang deficiency constitution, phlegm-dampness constitution, etc.), health status indicators (disease types, symptom severity, etc.), and health data change trends (weight change rate, blood pressure fluctuation value, etc.). The relative weights of each index are determined by AHP. Let the judgment matrix be A, where a ij represents the importance degree of index i relative to index j. The index weight vector W ` =(w ` 1,w ` 2,……,w ` n ) is obtained by calculating the maximum eigenvalue of matrix A and its corresponding eigenvector. Then, fuzzy evaluation is carried out for each index to determine its membership degree vector R i =(r i1 ,r i2 ,……,r im ) belonging to different evaluation levels (such as excellent, good, medium, poor), and the fuzzy relation matrix is constructed. Finally, through the fuzzy synthesis operation , where b j =max i min(w ` i, r ij ), the comprehensive evaluation result vector B is obtained, and based on this, the health assessment result of the user is determined, providing a basis for the generation of personalized solutions.

[0131] In an actual scenario, a survival analysis method can be introduced to evaluate the change of users' health risks over time. Let T be the time of occurrence of a health event (such as the occurrence of a disease, deterioration of health, etc.), and G = (g1, g2, ……, g p ) be the vector of influencing factors (including physical constitution type, health indicators, etc.). Through the Cox proportional hazards model h(t|G) = h0(t)exp(β1g1 + β2g2 + …… + β p g p ), where h(t|G) is the risk function at time t under the condition of given G, h0(t) is the baseline risk function, and β i is the regression coefficient. Calculate the health risk probability of users at different future time points according to the model, so as to plan preventive measures in advance in the health management plan.

[0132] As can be seen from the health status assessment method mentioned in the above embodiments, this method can collect and process heterogeneous data from multiple sources, and perform in-depth fusion analysis on them to obtain the physical constitution identification data of users, so as to obtain the physical constitution of users and perform health status analysis and evaluation on them to obtain accurate health assessment results, thus realizing the complete processing flow from data collection to health status assessment.

[0133] Corresponding to the health status assessment method provided in the foregoing embodiments, an embodiment of the present invention provides a health status assessment system, as Figure 9 shown, the system includes:

[0134] A health data collection unit 910, configured to collect physical constitution monitoring data of a user to be evaluated based on a preset multi-source heterogeneous database, and determine the health data of the user to be evaluated based on the semantic result of the physical constitution monitoring data;

[0135] A physical constitution identification processing unit 920, configured to obtain a feature vector corresponding to the health data, determine a weight matrix and a bias vector corresponding to the feature vector, and determine physical constitution identification data corresponding to the user to be evaluated by using the feature vector, the weight matrix and the bias vector;

[0136] A health data analysis unit 930, configured to perform density clustering calculation on the physical constitution identification data to obtain a health index clustering result of the user to be evaluated, and calculate a health index score of the user to be evaluated by using the health index clustering result;

[0137] An evaluation result generation unit 940, configured to determine the health risk level of the user to be evaluated according to the health index score, and generate a health evaluation result of the user to be evaluated based on a judgment matrix corresponding to the health risk level.

[0138] As can be seen from the health status assessment system mentioned in the above embodiments, the system can collect and process heterogeneous data from multiple sources, perform in-depth fusion analysis on it to obtain the user's physical constitution identification data, thereby obtaining the user's physical constitution and analyzing and evaluating the health status to obtain accurate health assessment results, thus realizing the complete processing process from data collection to health status assessment.

[0139] The health status assessment system provided by the embodiments of the present invention has the same implementation principle and technical effects as those of the foregoing embodiments of the health status assessment method. For the sake of brief description, for the parts not mentioned in the system embodiments, reference may be made to the corresponding content in the foregoing embodiments of the health status assessment method.

[0140] This embodiment also provides an electronic device, and the structural schematic diagram of the electronic device is as Figure 10 shown. The device includes a processor 101 and a memory 102; wherein, the memory 102 is used to store one or more computer instructions, and the one or more computer instructions are executed by the processor to implement the steps of the above health status assessment method.

[0141] Figure 10 The shown electronic device further includes a bus 103 and a communication interface 104, and the processor 101, the communication interface 104 and the memory 102 are connected through the bus 103.

[0142] Among them, the memory 102 may include a high-speed random access memory (RAM, Random Access Memory), and may also include a non-volatile memory, such as at least one disk memory. The bus 103 may be an ISA bus, a PCI bus or an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 10 only a bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0143] The communication interface 104 is used to connect to at least one user terminal and other network units through a network interface, and send the encapsulated IPv4 packet or IPv4 packet to the user terminal through the network interface.

[0144] The processor 101 may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 101 or the instructions in the form of software. The above-mentioned processor 101 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present disclosure. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present disclosure can be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 102, and the processor 101 reads the information in the memory 102 and combines its hardware to complete the steps of the method in the foregoing embodiments.

[0145] An embodiment of the present invention further provides a storage medium, on which a computer program is stored, and when the computer program is run by a processor, it executes the steps of the health status evaluation method in the foregoing embodiments.

[0146] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, equipment and methods can be implemented in other ways. The system embodiments described above are only illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For another example, 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 displayed or discussed coupling or direct coupling or communication connection to each other can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be in an electrical, mechanical or other form.

[0147] The unit described as a separation component may or may not be physically separated. The component displayed as a unit may or may not be a physical unit, that is, it may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0148] In addition, each functional unit in various embodiments of the present invention may be integrated in a processing unit, may exist separately as individual physical units, or two or more units may be integrated in one unit.

[0149] If the described function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0150] Finally, it should be noted that the above-described embodiments are only specific implementation manners of the present invention, used to illustrate the technical solutions of the present invention, and are not intended to limit it. The protection scope of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: any person skilled in the art within the technical scope disclosed by the present invention can still modify the technical solutions described in the foregoing embodiments, or can easily think of changes, or perform equivalent replacements on some of the technical features; and these modifications, changes, or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A health status assessment method, characterized in that, The method includes: A health data collection step: collecting physical monitoring data of a user to be evaluated based on a preset multi-source heterogeneous database, and determining the health data of the user to be evaluated based on the semantic result of the physical monitoring data; A physical constitution identification processing step: obtaining a feature vector corresponding to the health data, determining a weight matrix and a bias vector corresponding to the feature vector, and using the feature vector, the weight matrix, and the bias vector to determine the physical constitution identification data corresponding to the user to be evaluated; A health data analysis step: performing density clustering calculation on the physical constitution identification data to obtain a health index clustering result of the user to be evaluated, and calculating a health index score of the user to be evaluated using the health index clustering result; An evaluation result generation step: determining a health risk level of the user to be evaluated according to the health index score, and generating a health evaluation result of the user to be evaluated based on a judgment matrix corresponding to the health risk level; The physical constitution identification processing step includes: Obtaining text data corresponding to the health data, generating label sequence data corresponding to the text data based on an initialized bidirectional long short-term memory model and a conditional random field model, and using the label sequence data to determine the feature vector corresponding to the health data; Inputting the feature vector into a trained deep belief network, controlling the deep belief network to obtain modal data corresponding to the feature vector, and using a joint representation result between the modal data to determine the weight matrix and the bias vector corresponding to the feature vector; Determining hidden units corresponding to the deep belief network according to a preset activation function and using the feature vector, the weight matrix, and the bias vector, and outputting the physical constitution identification data corresponding to the user to be evaluated under the health data using the hidden units; The health data analysis step includes: Obtaining physical constitution type data and disease type data corresponding to the user to be evaluated based on the physical constitution identification data, and obtaining an association rule corresponding to the physical constitution type data and the disease type data; Performing density clustering calculation on the physical constitution identification data using the association rule, obtaining a feature distribution result between the physical constitution type data and the disease type data of the user to be evaluated, and determining the health index clustering result of the user to be evaluated based on the feature distribution result; Constructing a prediction model corresponding to the physical constitution type data and the disease type data according to an initialized Bayesian network, and calculating the health index score corresponding to the health index clustering result using the prediction model.

2. The health status assessment method according to claim 1, characterized in that The health data collection step includes: Determining a physical constitution questionnaire database, a four diagnostic database, and a health monitoring database based on the multi-source heterogeneous database; Obtaining the physical monitoring data corresponding to the user to be evaluated using the physical constitution questionnaire database, the four diagnostic database, and the health monitoring database. Construct a random forest corresponding to the physical fitness monitoring data, calculate the anomaly score of each data point in the random forest, and after performing an update process on the physical fitness monitoring data according to the anomaly score, obtain the semantic relationship corresponding to the physical fitness monitoring data; Use the semantic relationship to determine the semantic result corresponding to the physical fitness monitoring data, and determine the health data of the user to be evaluated according to the semantic result.

3. The health status assessment method according to claim 2, characterized in that Obtain the physical fitness monitoring data corresponding to the user to be evaluated by using the physical fitness questionnaire database, the four diagnostic databases, and the health monitoring database, including: Use the physical fitness questionnaire database to obtain the physical fitness questionnaire data corresponding to the user to be evaluated, and determine the first physical fitness monitoring data corresponding to the user to be evaluated according to the weight value and semantic result corresponding to the physical fitness questionnaire data; Use the four diagnostic databases to obtain the pulse condition data, tongue diagnosis data, voice data, and body temperature data corresponding to the user to be evaluated, and determine the second physical fitness monitoring data corresponding to the user to be evaluated according to the eigenvectors corresponding to the pulse condition data, the tongue diagnosis data, the voice data, and the body temperature data; Use the health monitoring database to obtain the health monitoring data and environmental data corresponding to the user to be evaluated, and determine the third physical fitness monitoring data corresponding to the user to be evaluated according to the health monitoring data and the environmental data; Determine the physical fitness monitoring data according to the first physical fitness monitoring data, the second physical fitness monitoring data, and the third physical fitness monitoring data.

4. The health status assessment method according to claim 1, wherein Construct a prediction model corresponding to the physical fitness type data and the disease type data according to the initialized Bayesian network, including: Obtain the first node set corresponding to the physical fitness type data and the second node set corresponding to the disease type data; Construct the corresponding conditional probability table of the Bayesian network based on the first node set and the second node set, and use the conditional probability table to determine the structural parameters of the Bayesian network; Construct the prediction model based on the Bayesian network according to the structural parameters.

5. The health status assessment method according to claim 1, wherein The evaluation result generation step includes: Obtain the window function corresponding to the health indicator clustering result, calculate the mean result of the health indicator scores under the window function, and determine the health risk level of the user to be evaluated according to the mean result; Construct an evaluation index strategy corresponding to the user to be evaluated based on the health indicator clustering result, and use the evaluation index strategy to determine the judgment matrix; Determine the health evaluation result corresponding to the user to be evaluated according to the judgment matrix.

6. The health status assessment method according to claim 5, wherein Determine the health evaluation result corresponding to the user to be evaluated according to the judgment matrix, including: Calculate the maximum eigenvalue of the judgment matrix and its corresponding eigenvector; Use the maximum eigenvalue and the corresponding eigenvector to obtain the index weight vector of the judgment matrix; Determine the membership degree vector corresponding to the index weight vector according to the health risk level, and construct the fuzzy relation matrix corresponding to the judgment matrix according to the membership degree vector; Generate an evaluation vector corresponding to the user to be evaluated according to the fuzzy relation matrix, and determine the health assessment result by using the evaluation vector.

7. A health status assessment system, characterized in that, The system includes: A health data acquisition unit, configured to collect physical fitness monitoring data of the user to be evaluated based on a preset multi-source heterogeneous database, and determine the health data of the user to be evaluated based on the semantic result of the physical fitness monitoring data; A physical constitution identification processing unit, configured to obtain a feature vector corresponding to the health data, determine a weight matrix and a bias vector corresponding to the feature vector, and determine physical constitution identification data corresponding to the user to be evaluated by using the feature vector, the weight matrix, and the bias vector; A health data analysis unit, configured to perform density clustering calculation on the physical constitution identification data to obtain a health index clustering result of the user to be evaluated, and calculate a health index score of the user to be evaluated by using the health index clustering result; An evaluation result generation unit, configured to determine a health risk level of the user to be evaluated according to the health index score, and generate a health assessment result of the user to be evaluated based on a judgment matrix corresponding to the health risk level; The physical constitution identification processing unit is further configured to: obtain text data corresponding to the health data, generate label sequence data corresponding to the text data based on an initialized bidirectional long short-term memory model and a conditional random field model, and determine the feature vector corresponding to the health data by using the label sequence data; input the feature vector into a trained deep belief network, control the deep belief network to obtain modal data corresponding to the feature vector, and determine the weight matrix and the bias vector corresponding to the feature vector by using a joint representation result between the modal data; determine a hidden unit corresponding to the deep belief network according to a preset activation function and by using the feature vector, the weight matrix, and the bias vector, and output the physical constitution identification data corresponding to the user to be evaluated under the health data by using the hidden unit; The health data analysis unit is further configured to: obtain physical constitution type data and disease type data corresponding to the user to be evaluated based on the physical constitution identification data, and obtain an association rule corresponding to the physical constitution type data and the disease type data; perform density clustering calculation on the physical constitution identification data by using the association rule, obtain a feature distribution result between the physical constitution type data and the disease type data of the user to be evaluated, and determine the health index clustering result of the user to be evaluated based on the feature distribution result; construct a prediction model corresponding to the physical constitution type data and the disease type data according to an initialized Bayesian network, and calculate the health index score corresponding to the health index clustering result by using the prediction model.

8. An electronic device, characterized in that, It includes a processor and a memory, where the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the steps of the health status assessment method according to any one of claims 1 to 6.

Images