Intelligent health management system

Abstract

The invention provides an intelligent health management system. The intelligent health management system is provided with a medical care terminal, a community terminal, a home terminal and a cloud server; the cloud server is in communication connection with the medical care terminal, the community terminal and the family terminal, and the family terminal is used for collecting and sending family health data of a target family; the cloud server is used for inputting the family health data into a chronic disease assessment model to obtain a chronic disease assessment result, generating a chronic disease onset risk report and sending the chronic disease onset risk report to the medical care terminal and the family terminal; the medical care terminal is used for determining diseased members according to the chronic disease onset risk report and sending a first notification to the community terminal; the community end is used for receiving the first notification and determining an emergency help-seeking object and sending an emergency help-seeking signal when receiving the second notification; and the medical care terminal is also used for obtaining and dynamically adjusting the treatment scheme according to the medicine taking history and the health change record of the diseased member, and sending a third notification to the home terminal, so that the chronic disease management capability is improved.

Description

Technical Field

[0001] This application relates to the field of medical health, and particularly to an intelligent health management system. Background Art

[0002] With the improvement of people's living standards, the incidence of chronic non-communicable diseases has been on the rise year by year. And effective health management is the key to preventing and controlling chronic diseases. However, most of the current domestic management models are not perfect, and it is difficult to achieve a relatively comprehensive effect through single-level health management. As a result, it is easy for chronic disease patients, their family members, community medical workers, and medical institutions to experience slow response and difficult management in the face of the needs of long-term chronic disease management, condition monitoring and tracking, health education and guidance, and doctor-patient communication. Summary of the Invention

[0003] The main purpose of the embodiments of this application is to propose an intelligent health management system. Through communication and interaction among the medical staff terminal, community terminal, family terminal, and cloud server, it can timely detect and handle the abnormal conditions of patients, improve the long-term management ability of chronic diseases, and shorten the communication distance among patients, community workers, and medical workers.

[0004] To achieve the above object, the first aspect of the embodiments of this application proposes an intelligent health management system, including: A medical staff terminal, a community terminal, a family terminal, and a cloud server; The cloud server is respectively communicatively connected to the medical staff terminal, the community terminal, and the family terminal; The family terminal is used to collect the family health data of the target family and send the family health data to the cloud server, and the family health data includes the physiological data, environmental data, lifestyle data, and family medical history data of each family member; The cloud server is used to receive the family health data and input the family health data into the chronic disease assessment model constructed based on the random forest algorithm to obtain the chronic disease assessment result; The cloud server is further used to periodically generate a chronic disease incidence risk report according to the chronic disease assessment result and send the chronic disease incidence risk report to the medical staff terminal and the family terminal, where the chronic disease incidence risk report includes the incidence risk assessment, risk factor analysis, and intervention suggestions of each family member.

[0005] The medical staff terminal is used to receive the chronic disease incidence risk report, determine the diseased members among each family member according to the chronic disease incidence risk report, and send a first notice to the community terminal, and the first notice is used to inform that the diseased members are determined as the community focus objects; The community side is used to receive the first notice, and in response to receiving the second notice from the cloud server, determine the community's concerned object as an emergency help-seeking object, and send out an emergency help signal. The second notice is used to inform that the physiological data of the community's concerned object has an abnormal situation; The medical care side is also used to regularly obtain the medication history and health change records of the diseased members from the cloud server, dynamically adjust the treatment plan of the diseased members according to the medication history and health change records, and send a third notice to the family side. The third notice is used to inform the diseased members to go to the hospital for a follow-up visit regularly.

[0006] Furthermore, in some embodiments, the chronic disease assessment model is constructed through the following steps: Obtain a chronic disease sample set, which includes a first physiological sample set, an environmental sample set, a first lifestyle sample set, and a family medical history sample set of multiple patients with chronic diseases; Extract features from the chronic disease sample set to obtain a chronic disease feature set; Divide the chronic disease feature set into a training set and a test set; According to the training set, initialize the decision tree network architecture and train the decision tree network architecture to obtain a chronic disease training model; According to the test set, optimize the parameters of the chronic disease training model to obtain a chronic disease assessment model.

[0007] Furthermore, in some embodiments, extracting features from the chronic disease sample set to obtain a chronic disease feature set includes: Based on the chi-square test method, extract features from the family medical history sample set to obtain a first feature set; Based on the principal component analysis method, extract features from the environmental sample set and the first lifestyle sample set respectively to obtain a second feature set and a third feature set; Based on the correlation analysis method, extract features from the first physiological sample set to obtain a fourth feature set; Merge the parameter dimensions of the first feature set, the second feature set, the third feature set, and the fourth feature set to obtain a chronic disease feature set.

[0008] Furthermore, in some embodiments, according to the training set, initialize the decision tree network architecture and train the decision tree network architecture to obtain a chronic disease training model, including: Through the Bootstrap sampling method, randomly draw samples from the training set with replacement multiple times to obtain multiple first training sets and a second training set. Among them, the second training set is the sample set that has never been selected and drawn in the training set; According to multiple first training sets, initialize multiple decision trees, and respectively split and grow each decision tree in a non-pruning strategy to obtain a decision tree network architecture; Input the second training set into the decision tree network architecture for multivariate prediction to obtain multiple prediction results; Perform regression training based on each prediction result to obtain a regression prediction architecture, and merge the regression prediction architecture with the decision tree network architecture to obtain a chronic disease training model; wherein, the regression prediction architecture is used to integrate each prediction result into an optimal prediction result.

[0009] Furthermore, in some embodiments, the process of training and growing the decision tree includes the following steps: Based on the information gain algorithm, calculate the information gain degree of each sample in the first training set, and select the sample with the maximum information gain degree as the splitting node; Determine the splitting value of the splitting node, and divide the first training set into multiple split subsets according to the splitting value; Respectively take the multiple split sets as the new first training set, and return to the step of calculating the information gain degree of each sample in the first training set based on the information gain algorithm and selecting the sample with the optimal information gain degree as the splitting node until the preset stop condition is met to obtain the decision tree.

[0010] Furthermore, in some embodiments, the cloud server is further used to obtain the physiological data, lifestyle data, and health change records of the diseased members, and input the physiological data, lifestyle data, and health change records into the health demand library for demand prediction according to the collaborative filtering method to obtain the target demand; The cloud server is further used to screen out content consultations that meet the target demand from the health article library and the health service library according to the target demand, and push the content consultations to the home end, wherein the content consultations include health education materials, health activity announcements, and health service hotlines.

[0011] Furthermore, in some embodiments, the health demand library is obtained through the following steps: Obtain the health sample sets and health demand sets of different users. The health sample sets include the second physiological sample set, the second lifestyle sample set, and the health change sample set. The health demand sets include multiple demand samples and the demand scores corresponding to the demand samples; Extract features from each health sample set to obtain multiple fifth feature sets; Extract features from each health demand set to obtain multiple sixth feature sets; Construct an association matrix between the fifth feature sets and the sixth feature sets of different users to obtain the health demand library.

[0012] Furthermore, in some embodiments, inputting the physiological data, lifestyle data, and health change records into the health demand library for demand prediction according to the collaborative filtering method to obtain the target demand includes: Calculate the first similarity between each user and the diseased members in the health needs library according to physiological data, lifestyle data, and health change records, and determine similar users in the health needs library based on the first similarity; Determine the same demand samples as similar demands in the health needs sets of each similar user, and determine the demand scores corresponding to each similar demand as similar scores; Perform a weighted average operation on each similar score of the same type according to the target similarity between the similar user and the diseased member to obtain multiple predicted scores; Among each predicted score, determine the similar demand corresponding to the predicted score with the highest score as the target demand.

[0013] Further, in some embodiments, calculating the first similarity between each user and the diseased members in the health needs library according to physiological data, lifestyle data, and health change records, and determining similar users in the health needs library based on the first similarity includes: Calculate the cosine similarity between each diseased member and each user in the health needs library respectively according to physiological data, lifestyle data, and health change records to obtain multiple first similarities; Determine the first similarity greater than or equal to the first preset threshold as the second similarity, and determine the user corresponding to the second similarity as a similar user in the health needs library.

[0014] Further, in some embodiments, the health management system further includes: A health detection device, which is communicatively connected to the home terminal; The health detection device is used to collect the physiological data and environmental data of each family member, and send the physiological data and environmental data to the home terminal.

[0015] The embodiments of the present application have the following beneficial effects: by providing a medical care terminal, a community terminal, a family terminal, and a cloud server; the cloud server is respectively communicatively connected to the medical care terminal, the community terminal, and the family terminal; the family terminal is used to collect the family health data of the target family and send the family health data to the cloud server, and the family health data includes the physiological data, environmental data, lifestyle data, and family medical history data of each family member; the cloud server is used to receive the family health data and input the family health data into a chronic disease assessment model constructed based on the random forest algorithm to obtain a chronic disease assessment result; the cloud server is also used to periodically generate a chronic disease onset risk report according to the chronic disease assessment result and send the chronic disease onset risk report to the medical care terminal and the family terminal, wherein the chronic disease onset risk report includes the onset risk assessment, risk factor analysis, and intervention suggestions of each family member, the medical care terminal is used to receive the chronic disease onset risk report and determine the diseased members among each family member according to the chronic disease onset risk report, and send a first notice to the community terminal, and the first notice is used to inform that the diseased members are determined as community focus objects; the community terminal is used to receive the first notice and, in response to receiving a second notice from the cloud server, determine the community focus objects as emergency assistance objects and issue an emergency assistance signal, and the second notice is used to inform that the physiological data of the community focus objects has an abnormal situation; the medical care terminal is also used to regularly obtain the medication history and health change records of the diseased members from the cloud server, dynamically adjust the treatment plan of the diseased members according to the medication history and health change records, and send a third notice to the family terminal, and the third notice is used to inform the diseased members to go to the hospital for a follow-up visit regularly. Furthermore, through the communication and interaction among the medical care terminal, the community terminal, the family terminal, and the cloud server, the abnormal situations of patients can be discovered and processed in a timely manner, the long-term management ability of chronic diseases is improved, and the communication distance among patients, community workers, and medical workers is shortened. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is an architecture diagram of a health management system provided by some embodiments of the present application; Figure 2 is a flowchart of constructing a chronic disease assessment model provided by some embodiments of the present application; Figure 3 is provided by some embodiments of the present application Figure 2 flowchart of step S202 in; Figure 4 is provided by some embodiments of the present application Figure 2 flowchart of step S204 in; Figure 5 is a flowchart of the training and growth of a decision tree provided by some embodiments of the present application; Figure 6 is a flowchart of constructing a health needs library provided by some embodiments of the present application; Figure 7 It is a flowchart for demand prediction of the health demand library provided by some embodiments of the present application; Figure 8 It is provided by some embodiments of the present application Figure 7 The flowchart of step S701 in Figure 9 It is a schematic flowchart of the health management method provided by some embodiments of the present application; Figure 10 It is a schematic hardware structure diagram of an electronic device provided by some embodiments of the present application. Detailed implementation manners

[0017] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application 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 only used to explain the present application, and are not used to limit the present application.

[0018] In the description of the present application, it should be understood that for the orientation description, such as the orientation or positional relationship indicated by up, down, front, back, left, right, etc. is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present application.

[0019] It should also be noted that in the description of the present application, the meaning of several is more than one, the meaning of multiple is more than two, greater than, less than, exceeding, etc. are understood as not including the present number, and above, below, within, etc. are understood as including the present number. If there is a description of first and second, it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or implicitly indicating the sequence relationship of the indicated technical features.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application, and are not intended to limit the present application.

[0021] In the description of the present application, the descriptions with reference to terms such as "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0022] With the improvement of people's living standards, the incidence of chronic non-communicable diseases has shown an increasing trend year by year. And effective health management is the key to preventing and controlling chronic diseases. However, most of the current domestic management models are not perfect, and it is very difficult to achieve a relatively comprehensive effect only through single-level health management. As a result, it is easy for chronic disease patients, their family members, community medical workers, and medical institutions to experience slow response speed and management difficulties in the face of the needs of long-term management of chronic diseases, condition monitoring and tracking, health education and guidance, and doctor-patient communication.

[0023] Based on this, the embodiments of the present application provide an intelligent health management system. Through the communication and interaction among the medical staff terminal, the community terminal, the family terminal, and the cloud server, it can timely discover and handle the abnormal conditions of patients, improve the long-term management ability of chronic diseases, and shorten the communication distance among patients, community workers, and medical workers.

[0024] An intelligent health management system provided by the embodiments of the present application will be specifically described through the following embodiments.

[0025] Refer to Figure 1 As shown in Figure 1 is the architecture diagram of the health management system provided by some embodiments of the present application. The intelligent health management system is provided with a medical staff terminal 101, a community terminal 102, a family terminal 103, a cloud server 104, and a health detection device 105; the cloud server 104 is respectively communicatively connected to the medical staff terminal 101, the community terminal 102, and the family terminal 103, and the health detection device 105 is communicatively connected to the family terminal 103; Among them, the health detection device 105 is used to collect the physiological data and environmental data of each family member, and send the physiological data and environmental data to the family terminal 103.

[0026] The family terminal 103 is used to collect the family health data of the target family, and send the family health data to the cloud server 104. The family health data includes the physiological data, environmental data, lifestyle data, and family medical history data of each family member.

[0027] The cloud server 104 is used to receive home health data and input the home health data into a chronic disease assessment model constructed based on the random forest algorithm to obtain a chronic disease assessment result.

[0028] The cloud server 104 is also used to periodically generate a chronic disease onset risk report according to the chronic disease assessment result and send the chronic disease onset risk report to the medical care terminal 101 and the home terminal 103.

[0029] Among them, the chronic disease onset risk report includes the onset risk assessment, risk factor analysis, and intervention suggestions for each family member.

[0030] The medical care terminal 101 is used to receive the chronic disease onset risk report, determine the diseased members among each family member according to the chronic disease onset risk report, and send a first notice to the community terminal 102. The first notice is used to inform that the diseased members are determined as community focus objects; The community terminal 102 is used to receive the first notice, determine the community focus objects as emergency assistance objects in response to receiving a second notice from the cloud server 104, and send an emergency assistance signal. The second notice is used to inform that the physiological data of the community focus objects has abnormal conditions; The medical care terminal 101 is also used to regularly obtain the medication history and health change records of the diseased members from the cloud server 104, dynamically adjust the treatment plan of the diseased members according to the medication history and health change records, and send a third notice to the home terminal 103. The third notice is used to inform the diseased members to go to the hospital for a follow-up visit regularly.

[0031] Furthermore, the cloud server 104 is also used to obtain the physiological data, lifestyle data, and health change records of the diseased members, and input the physiological data, lifestyle data, and health change records into a health needs library for demand prediction according to the collaborative filtering method to obtain target needs.

[0032] Moreover, the cloud server 104 is also used to screen out content consultations that meet the target needs from a health article library and a health service library according to the target needs and push the content consultations to the home terminal 103. Among them, the content consultations include health education materials, health activity announcements, and health service hotlines.

[0033] Refer to Figure 2 shown Figure 2 is a flowchart of constructing a chronic disease assessment model provided by some embodiments of the present application. The method for constructing the chronic disease assessment model may include but is not limited to steps S201 to S205.

[0034] Step S201: Obtain a chronic disease sample set.

[0035] Among them, the chronic disease sample set includes the first physiological sample set, environmental sample set, first lifestyle sample set, and family medical history sample set of multiple patients with chronic diseases.

[0036] Step S202: Extract features from the chronic disease sample set to obtain a chronic disease feature set.

[0037] Specifically, in the process of extracting features from the chronic disease sample set, first, strict preprocessing of the original data is carried out, including data cleaning, missing value filling, and standardization processing to ensure data quality and consistency. Then, multi-dimensional information such as heart rate, blood pressure, blood biochemical indicators, life stress, and behavior habits is extracted from the patient's physiological indicators, lifestyle, and environmental factors. Variables closely related to chronic diseases are screened out through domain knowledge and statistical analysis. Subsequently, through data dimensionality reduction methods, the initially extracted high-dimensional features are optimized and integrated to construct a chronic disease feature set that has both high discrimination and rich information. This feature set not only comprehensively reflects the biological characteristics and clinical manifestations of the object members but also provides a solid data basis for subsequent model training, disease risk assessment, and the formulation of personalized treatment plans.

[0038] Step S203: Divide the chronic disease feature set into a training set and a test set.

[0039] Specifically, after constructing the chronic disease feature set, the next step is to divide this feature set into a training set and a test set for subsequent model construction and evaluation. Usually, a random sampling method is adopted to divide the overall data according to a preset ratio (such as 70% for training and 30% for testing) to ensure that the training set fully reflects the overall distribution of the data, while the test set is used as an independent data set to evaluate the generalization ability and prediction accuracy of the model. In addition, to avoid introducing biases due to class imbalance, a stratified sampling method can be adopted to keep the proportion of each class consistent in the training set and the test set, thereby improving the accuracy of model evaluation. Through this data set division strategy, not only sufficient training samples are provided for the model, but also the robustness and reliability of the model in real application scenarios can be objectively detected during the subsequent evaluation process.

[0040] Step S204: Initialize the decision tree network architecture according to the training set and train the decision tree network architecture to obtain a chronic disease training model.

[0041] Specifically, initialize the decision tree network architecture, select an appropriate algorithm (such as CART, ID3, or C4.5) and parameter settings (such as maximum depth, minimum sample split count, etc.) to adapt to the characteristics of chronic disease data. Then, use the training set to train the decision tree model. The model recursively selects the best features for data partitioning and constructs a tree structure until the stopping condition is met (such as reaching the maximum depth or node purity). After training, the obtained chronic disease training model can be used to predict new samples, assisting in clinical decision-making and the formulation of personalized treatment plans.

[0042] Step S205: According to the test set, optimize the parameters of the chronic disease training model to obtain a chronic disease evaluation model.

[0043] Specifically, during the process of optimizing the parameters of the chronic disease training model using the test set, first, a set of possible hyperparameter combinations need to be defined, such as the maximum depth of the decision tree, the minimum sample split count, and the splitting criterion, etc. Then, adopt the cross-validation method to evaluate the performance of each parameter combination on the test set and select the parameter combination that makes the model performance optimal. Through this process, the parameters of the training model are optimized, and finally, a chronic disease evaluation model is formed.

[0044] In actual operation, methods such as GridSearchCV or RandomizedSearchCV can be used to systematically optimize the parameters. For example, GridSearchCV finds the best parameters by exhaustively enumerating all possible parameter combinations, while RandomizedSearchCV randomly selects combinations in the parameter space for evaluation.

[0045] Refer to Figure 3 as shown Figure 3 is the flowchart of step S202 provided by some embodiments of the present application, and this method may include but is not limited to steps S301 to S304. Figure 2

[0046] Step S301: Based on the chi-square test method, extract features from the family history sample set to obtain a first feature set.

[0047] In a possible implementation manner, when extracting features from the family history sample set, through the chi-square test, the degree of association between each feature in the family history and the occurrence of the disease can be determined, so as to screen out the features with strong correlation with the disease. Specifically, the chi-square test calculates the chi-square statistic by comparing the observed frequency and the expected frequency, and judges the relevance between the feature and the disease according to its significance level.

[0048] Step S302: Based on the principal component analysis method, extract features from the environmental sample set and the first lifestyle sample set respectively to obtain a second feature set and a third feature set.

[0049] Among them, the environmental sample set usually contains multiple variables, such as air quality indicators, water quality parameters, noise levels, etc. Through the PCA principal component analysis method, these variables with strong correlations can be combined into a few principal components. Each principal component represents a linear combination of the original variables and reflects the main variation directions in the data. For example, PCA is used for pattern recognition of environmental mixtures. By reducing the dimensionality of exposure data of multiple pollutants, the main exposure patterns are identified, thus providing valuable information for environmental health research.

[0050] Furthermore, the first lifestyle sample set may include multiple indicators such as eating habits, exercise frequency, sleep quality, etc. Applying PCA, these indicators can be integrated into a few principal components to simplify the data structure and facilitate subsequent analysis and modeling.

[0051] Step S303: Based on the correlation analysis method, perform feature extraction on the first physiological sample set to obtain the fourth feature set.

[0052] When performing feature extraction on the first physiological sample set, the correlation analysis method is adopted to identify and screen out physiological indicators that are closely related to the target variable and have a large amount of information. By calculating the Pearson or Spearman correlation coefficients between various indicators, a correlation matrix is constructed, and variable pairs with high correlations (for example, the correlation coefficient is greater than 0.7) are identified from it. For these highly correlated indicators, according to the model objective, the part with the largest amount of information is selected for retention, redundant information is removed or merged, thus effectively reducing the interference of multicollinearity on subsequent model building. Finally, through this process, a set of representative, information-rich and dimension-reduced core physiological features is extracted, which constitutes the fourth feature set and provides a solid data basis for subsequent risk assessment and prediction modeling of chronic diseases.

[0053] Step S304: Perform parameter dimension merging on the first feature set, the second feature set, the third feature set and the fourth feature set to obtain the chronic disease feature set.

[0054] After integrating the first feature set, the second feature set, the third feature set and the fourth feature set, a comprehensive chronic disease feature set can be constructed. In this process, through parameter dimension merging technology, features from different dimensions such as family history, environmental factors, lifestyle and physiological indicators are spliced or weighted and fused. Such integration not only retains the key information in each feature set, but also effectively reduces the influence of redundant data and noise, thus constructing a feature set that comprehensively reflects the multi-factor influence of chronic diseases and provides a solid data basis for subsequent model training and risk assessment.

[0055] Refer to Figure 4 as shown Figure 4Provided by some embodiments of the present application Figure 2 It is a flowchart of step S204 in Figure 2 . The method may include but is not limited to steps S401 to S404.

[0056] Step S401: By the Bootstrap sampling method, samples are randomly drawn from the training set with replacement multiple times to obtain multiple first training sets and second training sets.

[0057] Among them, the second training set is a sample set that has never been selected and drawn from the training set.

[0058] Specifically, by the Bootstrap sampling method, multiple random samplings are performed from the original training set in a sampling-with-replacement manner. The number of samples drawn each time is usually the same as that of the original training set, thereby constructing multiple independent first training sets. This method not only ensures that each first training set can fully represent the distribution of the overall data. These first training sets obtained by repeated sampling multiple times provide a rich and diverse data source for the subsequent training of the model, enabling the constructed model to effectively reduce variance, improve robustness and generalization ability, and also providing a solid foundation for algorithms such as ensemble learning.

[0059] Step S402: Initialize multiple decision trees according to the multiple first training sets, and split and grow each decision tree in a non-pruning strategy manner to obtain a decision tree network architecture.

[0060] Specifically, initialize a decision tree model for each first training set. Each decision tree starts from a root node, and at this time, the decision tree has not undergone any splitting. Next, for each decision tree, adopt a non-pruning strategy, that is, allow the decision tree to grow fully until a preset stop condition is met, such as reaching the maximum depth or the number of samples in the node is less than a certain threshold. During the growth process, each decision tree randomly selects a part of the features at each node, and then selects the optimal feature and splitting point for splitting according to the information gain index of these features. In this way, each decision tree gradually forms its own structure, and finally constitutes a decision tree network architecture.

[0061] The above method ensures that each tree can capture the subtle differences in the training data. Although a single tree may overfit, the generalization ability of the model is improved by integrating multiple such decision trees. Finally, these independently trained decision trees together constitute a powerful decision tree network architecture, providing a solid foundation for the prediction and analysis of chronic diseases.

[0062] Step S403: Input the second training set into the decision tree network architecture for multivariate prediction to obtain multiple prediction results.

[0063] Specifically, when the data of the second training set is input into the decision tree network architecture, each decision tree will make predictions on the data according to its own structure and splitting rules. Since each decision tree in the decision tree network architecture grows independently, their prediction results for the second training set will also be different. These different prediction results together constitute multiple prediction results, providing rich information for subsequent analysis and decision-making. In this way, the powerful computing ability of the decision tree network architecture can be utilized to perform multivariate prediction on the data in the second training set, so as to better understand and solve complex problems.

[0064] Step S404: Perform regression training based on each prediction result to obtain a regression prediction architecture, and merge the regression prediction architecture with the decision tree network architecture to obtain a chronic disease training model.

[0065] Among them, the regression prediction architecture is used to integrate each prediction result into an optimal prediction result.

[0066] Specifically, perform regression training based on each prediction result to construct a regression prediction architecture. The purpose of regression training is to further improve the prediction accuracy of the model by learning the relationship between the prediction result and the actual target value. In this process, various linear regression algorithms can be used. Through regression training, a regression prediction architecture that can finely adjust the prediction result is obtained.

[0067] Finally, merge the regression prediction architecture with the decision tree network architecture to obtain a chronic disease training model. The merged model combines the multivariate prediction ability of the decision tree network architecture and the fine adjustment ability of the regression prediction architecture, and can more accurately predict the risk of chronic diseases. This chronic disease training model can be used in actual disease prediction and health management to provide personalized risk assessment and intervention suggestions for patients. In this way, the advantages of the decision tree network architecture and the regression prediction architecture can be fully utilized to construct an efficient and accurate chronic disease prediction model.

[0068] Refer to Figure 5 as shown Figure 5 is a flowchart of the training and growth of a decision tree provided by some embodiments of the present application. The method may include but is not limited to steps S501 to S503.

[0069] Step S501: Based on the information gain algorithm, calculate the information gain degree of each sample in the first training set, and select the sample with the maximum information gain degree as the splitting node.

[0070] Specifically, the information gain degree of each sample is calculated based on the information gain algorithm. First, the entropy of the parent node is calculated. Entropy is an indicator to measure the purity of the data set. The higher the entropy value, the higher the degree of chaos of the data set. Then, for each feature, it is used as a potential splitting point to divide the data set into several subsets, and the entropy of each subset is calculated. Next, according to the change in entropy values before and after splitting, the information gain of this feature is calculated. The information gain reflects the reduction in uncertainty that can be achieved by splitting using this feature. Finally, the information gains of all features are compared, and the feature with the maximum information gain degree is selected as the splitting node. In this way, it can be ensured that during the construction of the decision tree, each split can maximize the purity of the data set, thereby improving the classification or regression performance of the decision tree.

[0071] Step S502: Determine the splitting value of the splitting node, and divide the first training set into multiple splitting subsets according to the splitting value.

[0072] In a possible embodiment, assume a first training set that contains information such as the age of patients, cholesterol levels, and whether they have hyperlipidemia. First, a feature needs to be selected as the splitting node, and then the splitting value of this feature is determined to divide the data set into as pure subsets as possible. Taking the selection of cholesterol level as the splitting feature as an example, the optimal splitting value is determined by calculating the information gain or Gini impurity at different cholesterol level thresholds. The possible splitting values to be considered include 150mg / dL, 200mg / dL, and 250mg / dL, etc. By calculating the information gain of each possible splitting value, the splitting value that leads to the greatest increase in data purity can be selected. For example, if 200mg / dL is used as the splitting value, 80% of the patients with cholesterol levels lower than or equal to 200mg / dL do not have hyperlipidemia, while 90% of the patients with cholesterol levels higher than 200mg / dL have hyperlipidemia. Then this splitting value has a relatively high information gain. Select this splitting value as the splitting node, and then divide the first training set into two subsets: one subset contains patients with cholesterol levels lower than or equal to 200mg / dL, and the other subset contains patients with cholesterol levels higher than 200mg / dL. Each subset can then be further split to construct a more detailed decision tree structure and improve the prediction ability of the model.

[0073] Step S503: Respectively regard the multiple splitting sets as new first training sets, and return to the step of calculating the information gain degree of each sample in the first training set based on the information gain algorithm and selecting the sample with the optimal information gain degree as the splitting node until the preset stop condition is met, and a decision tree is obtained.

[0074] Specifically, the multiple split sets are respectively used as new first training sets again, and the above step S601 is returned until a preset stop condition is met to obtain a decision tree, where the preset stop condition includes but is not limited to: the decision tree reaches the maximum tree depth, the information gain degree is lower than the pre-threshold, etc.

[0075] Refer to Figure 6 as shown Figure 6 is a flowchart of constructing a health needs library provided by some embodiments of the present application. The method may include but is not limited to steps S601 to S604.

[0076] Step S601: Obtain the health sample sets and health needs sets of different users.

[0077] Among them, the health sample set includes a second physiological sample set, a second lifestyle sample set, and a health change sample set, and the health needs set includes multiple demand samples and demand scores corresponding to the demand samples.

[0078] Step S602: Extract features from each health sample set to obtain multiple fifth feature sets.

[0079] Specifically, the features extracted from each sample set include data in multiple dimensions such as physiological indicators, lifestyle factors, and health change samples. For example, physiological indicators can cover blood pressure, blood sugar, blood lipid levels, etc., and lifestyle factors include eating habits, exercise frequency, smoking and drinking situations, etc. By integrating this information, multiple fifth feature sets can be obtained, and each feature set contains detailed features closely related to the health status. These feature sets will serve as important inputs for model training, enabling more accurate identification and prediction of health risk factors, thereby providing a scientific basis for individualized health management.

[0080] Step S603: Extract features from each health needs set to obtain multiple sixth feature sets.

[0081] Specifically, the features extracted from each health needs set can be exercise rehabilitation services, health lectures, health check-ups, and psychological counseling services. By integrating this information, multiple sixth feature sets can be obtained, and each feature set contains detailed features closely related to the health needs.

[0082] Step S604: Construct an association matrix between the fifth feature sets and the sixth feature sets of different users to obtain a health needs library.

[0083] In the process of constructing a health needs library, it is first necessary to deeply analyze the fifth feature set and the sixth feature set of different users. The fifth feature set usually contains information about users' health conditions, such as physiological indicators, medical history, lifestyle, etc., while the sixth feature set covers users' needs and preferences for health services, such as service types, service frequencies, satisfaction levels, etc. By constructing an association matrix between these two types of feature sets, the internal relationship between users' health characteristics and service needs can be revealed.

[0084] Specifically, the steps for constructing the association matrix are as follows: First, standardize the fifth feature set and the sixth feature set to eliminate the dimensional differences between different features. Then, use the method of correlation coefficients to calculate the correlation between each feature in the fifth feature set and each feature in the sixth feature set. Next, fill these correlation indicators into the association matrix to form a complete association matrix. This matrix can intuitively display the degree of association between different health characteristics and service needs.

[0085] Based on this association matrix, a health needs library can be further constructed. By analyzing the high-correlation feature pairs in the association matrix, the most suitable health service combinations can be recommended for different user groups. For example, for users with specific health risks, according to the key indicators in their fifth feature set, the corresponding service needs in the sixth feature set can be matched to achieve personalized health service recommendations. This method not only improves the pertinence and effectiveness of health services but also provides a scientific basis for health management.

[0086] Refer to Figure 7 as shown Figure 7 is a flowchart for demand prediction of the health needs library provided by some embodiments of the present application. The method may include but is not limited to steps S701 to S704.

[0087] Step S701: Calculate the first similarity between each user in the health needs library and the diseased members according to physiological data, lifestyle data, and health change records, and determine similar users in the health needs library according to the first similarity.

[0088] Specifically, calculate the first similarity between each user in the health needs library and the diseased members according to physiological data, lifestyle data, and health change records, and determine similar users in the health needs library according to the first similarity.

[0089] Step S702: Determine the same demand samples as similar demands in the health needs sets of each similar user, and determine the demand scores corresponding to each similar demand as similar scores.

[0090] For example, if the health needs of one similar user mainly include Need A and Need B, and the health needs of another similar user also mainly include Need A and Need B, then Need A and Need B are determined as similar needs. The need scores corresponding to Need A are obtained from the health needs of the two similar users respectively, and the need scores corresponding to Need B are obtained from the health needs of the two similar users respectively. These four need scores are determined as similar scores.

[0091] Step S703: According to the target similarity between the similar user and the diseased member, weighted average operations are respectively performed on each type of similar scores to obtain multiple predicted scores.

[0092] Exemplarily, assume that the target similarity between similar user X and the diseased member is 0.8, the target similarity between similar user Y and the diseased member is 0.7, and there are similar needs A and similar needs B. The similar score of similar need A in similar user X is 9, the similar score of similar need A in similar user Y is 8, the similar score of similar need B in similar user X is 7, and the similar score of similar need B in similar user Y is 8.

[0093] Therefore, weighted average operations are performed on the similar scores of the same type (i.e., the similar scores of similar need A in similar user X and similar user Y respectively) to obtain the predicted score corresponding to similar need A ; weighted average operations are performed on the similar scores of the same type (i.e., the similar scores of similar need B in similar user X and similar user Y respectively) to obtain the predicted score corresponding to similar need B .

[0094] Step S704: Among the predicted scores, the similar need corresponding to the predicted score with the highest value is determined as the target need.

[0095] Taking the predicted score 8.53 corresponding to similar need A and the predicted score 7.47 corresponding to similar need B as an example, it can be seen that 8.53 is the predicted score with the highest value. Therefore, the similar need A corresponding to the predicted score 8.53 is determined as the target need.

[0096] Refer to Figure 8 as shown Figure 8 is the flowchart of step S701 provided by some embodiments of the present application. The cache garbage collection method may include but is not limited to steps S801 to S802. Figure 7 Step S801: According to the physiological data, lifestyle data, and health change records, calculate the cosine similarity between the diseased member and each user in the health needs library respectively to obtain multiple first similarities.

[0097]

[0098] ​Specifically, physiological data covers indicators such as blood pressure, blood sugar, and heart rate. Lifestyle data involves aspects such as exercise frequency, eating habits, and work and rest patterns. Health change records include disease symptoms, treatment progress, etc. By quantifying these data and converting them into vector form, then using the cosine similarity formula, that is, the dot product of vectors divided by the product of the vector norms, to measure the similarity degree between the diseased members and each user in the health demand library in the multi-dimensional data space. For example, for the diseased member A, represent his physiological data, lifestyle data, and health change records as vectors a, b, and c respectively, and the users in the health demand library correspond to their respective vectors. By calculating the cosine value of the angle between the vectors, the similarity degree of different users and the diseased member in terms of health characteristics can be obtained. The closer the similarity value is to 1, the higher the similarity, and vice versa. Through this series of calculations, users similar to the diseased member in terms of their health status can be found, and targeted health service suggestions can be provided accordingly.

[0099] Step S802: Determine the first similarity that is greater than or equal to the first preset threshold as the second similarity, and determine the similar users as the users corresponding to the second similarity in the health demand library.

[0100] Specifically, first, the dynamically generated first similarity is compared with the preset threshold parameter. When the similarity value reaches or exceeds the first preset threshold (usually set as an interval value in the range of 0.75 - 0.85 according to the service scenario), then the feature enhancement mechanism is triggered, and the value is calibrated to a more discriminative second similarity through a weighted algorithm. This process will focus on enhancing the core feature dimensions in the user's health profile, including but not limited to key parameters such as body mass index, chronic disease types, and exercise habits. After calibration, a hierarchical search is implemented in the full user pool of the health demand library, and the user data with the top 10% in the second similarity ranking is preferentially extracted, and a set of similar users is generated through cluster analysis. This set will serve as the benchmark group for personalized service recommendations, providing data support for the formulation of subsequent health intervention plans, precise medical resource matching, and the construction of health communities, and ultimately realizing the "one-size-fits-one" health management service.

[0101] In a second aspect, the embodiments of the present application further provide a health management method, which is applied to a health management system. The health management system is provided with a medical staff terminal, a community terminal, a family terminal, and a cloud server; the cloud server is respectively communicatively connected with the medical staff terminal, the community terminal, and the family terminal. Refer to Figure 9 , Figure 9 is a schematic flowchart of the health management method provided by some embodiments of the present application. The health management method may include but is not limited to steps S901 to S905: Step S901: The cloud server receives the home health data sent by the home end, and inputs the home health data into the chronic disease assessment model constructed based on the random forest algorithm to obtain the chronic disease assessment result.

[0102] Among them, the home health data includes the physiological data, environmental data, lifestyle data and family medical history data of each family member Step S902: The cloud server periodically generates a chronic disease onset risk report according to the chronic disease assessment result, and sends the chronic disease onset risk report to the medical care end and the home end.

[0103] Among them, the chronic disease onset risk report includes the onset risk assessment, risk factor analysis and intervention suggestions of each family member.

[0104] Step S903: The medical care end receives the chronic disease onset risk report, determines the diseased members among each family member according to the chronic disease onset risk report, and sends a first notice to the community end.

[0105] Among them, the first notice is used to inform that the diseased members are determined as the community focus objects.

[0106] Step S904: The community end receives the first notice, and in response to receiving a second notice from the cloud server, determines the community focus object as an emergency assistance object and issues an emergency assistance signal. The second notice is used to inform that the physiological data of the community focus object has an abnormal condition.

[0107] Step S905: The medical care end regularly obtains the medication history and health change records of the diseased members from the cloud server, dynamically adjusts the treatment plan of the diseased members according to the medication history and health change records, and sends a third notice to the home end. The third notice is used to inform the diseased members to go to the hospital for a follow-up visit regularly.

[0108] The above-mentioned health management method and the above-mentioned health management system 100 are based on the same inventive concept. The above process describes the health management method of the embodiments of the present application. The cloud server receives the home health data sent by the home terminal, and inputs the home health data into the chronic disease assessment model constructed based on the random forest algorithm to obtain the chronic disease assessment result. Then, the cloud server periodically generates a chronic disease onset risk report according to the chronic disease assessment result, and sends the chronic disease onset risk report to the medical care terminal and the home terminal. Next, the medical care terminal receives the chronic disease onset risk report, determines the diseased members among each family member according to the chronic disease onset risk report, and sends a first notice to the community terminal. The first notice is used to inform that the diseased members are determined as the community focus objects. Next, the community terminal receives the first notice, and in response to receiving a second notice from the cloud server, determines the community focus objects as the emergency assistance objects, and issues an emergency assistance signal. The second notice is used to inform that the physiological data of the community focus objects has abnormal conditions. Next, the medical care terminal regularly obtains the medication history and health change records of the diseased members from the cloud server, dynamically adjusts the treatment plan of the diseased members according to the medication history and health change records, and sends a third notice to the home terminal. The third notice is used to inform the diseased members to go to the hospital for a follow-up visit regularly. Furthermore, through the communication and interaction among the medical care terminal, the community terminal, the home terminal and the cloud server, the abnormal conditions of the patients can be discovered and processed in time, the long-term management ability of chronic diseases is improved, and the communication distance among the patients, community workers and medical workers is shortened.

[0109] An embodiment of the present application also provides an electronic device. The electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above-mentioned health management method is implemented. The electronic device can be any intelligent terminal including a mobile phone, a tablet computer, an in-vehicle computer, etc.

[0110] Please refer to Figure 10 , Figure 10 which is a schematic diagram of the hardware structure of the electronic device provided by some embodiments of the present application. The electronic device includes: A processor 1001, which can be implemented by using a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the health management method provided by the embodiments of the present application; The memory 1002 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 1002 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1002 and are called by the processor 1001 to execute the health management method provided in the embodiments of this application; The input / output interface 1003 is used to implement information input and output; The communication interface 1004 is used to implement communication interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or through wireless means (such as mobile network, WIFI, Bluetooth, etc.); The bus 1005 transmits information between the various components of the device (such as the processor 1001, the memory 1002, the input / output interface 1003, and the communication interface 1004); Among them, the processor 1001, the memory 1002, the input / output interface 1003, and the communication interface 1004 are communicatively connected to each other inside the device through the bus 1005.

[0111] The embodiments of this application also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it is the health management method provided in the embodiments of this application.

[0112] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include high-speed random access memory, and can also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory optionally includes a memory remotely provided with respect to the processor, and these remote memories can be connected to the processor through a network. Examples of the above networks include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0113] The embodiments described in the embodiments of this application are for more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. Those skilled in the art know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are equally applicable to similar technical problems.

[0114] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown, or combine certain steps, or different steps.

[0115] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0116] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices can be implemented as software, firmware, hardware and their appropriate combinations.

[0117] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that comprises a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0118] It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally means that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or a similar expression means any combination of these items, including any combination of single item (one) or plural items (ones). For example, at least one (one) of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0119] In several embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of the above units is only a logical function division. In actual implementation, there may be other division methods. For 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 couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of devices or units can be in electrical, mechanical or other forms.

[0120] The units described above as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be 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.

[0121] In addition, each functional unit in various embodiments of the present application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0122] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-accessible storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or 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 multiple instructions for causing 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 in various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store programs.

[0123] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of rights of the embodiments of the present application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of rights of the embodiments of the present application.

Claims

1. An intelligent health management system, characterized in that: include: Medical end, community end, home end and cloud server; The cloud server is respectively connected to the medical end, the community end and the family end in communication; The family terminal is used to collect family health data of the target family and send the family health data to the cloud server, wherein the family health data includes physiological data, environmental data, lifestyle data and family medical history data of each family member; The cloud server is used to receive the family health data, and input the family health data into a chronic disease assessment model constructed based on a random forest algorithm to obtain a chronic disease assessment result; The cloud server is also used to periodically generate a chronic disease risk report based on the chronic disease assessment results, and send the chronic disease risk report to the medical and nursing end and the family end; wherein the chronic disease risk report includes the risk assessment, risk factor analysis and intervention suggestions of each family member; The medical end is used to receive the chronic disease risk report, determine the sick member among the family members according to the chronic disease risk report, and send a first notification to the community end, wherein the first notification is used to inform that the sick member is determined as a community concern object; The community terminal is used to receive the first notification, and in response to receiving a second notification from the cloud server, determine the community concerned object as an emergency help object, and send an emergency help signal, wherein the second notification is used to inform the community concerned object that the physiological data is abnormal; The medical end is also used to periodically obtain the medication history and health change records of the sick member from the cloud server, dynamically adjust the treatment plan of the sick member according to the medication history and the health change records, and send a third notification to the family end, wherein the third notification is used to inform the sick member to go to the hospital for follow-up at a regular time.

2. The health management system according to claim 1, characterized in that: The chronic disease assessment model is constructed by the following steps: Acquire a chronic disease sample set, wherein the chronic disease sample set includes a first physiological sample set, an environmental sample set, a first lifestyle sample set, and a family medical history sample set of a plurality of patients suffering from chronic diseases; Extracting features from the chronic disease sample set to obtain a chronic disease feature set; Dividing the chronic disease feature set into a training set and a test set; Initializing a decision tree network architecture according to the training set, and training the decision tree network architecture to obtain a chronic disease training model; According to the test set, the parameters of the chronic disease training model are tuned to obtain the chronic disease assessment model.

3. The health management system according to claim 2, characterized in that: The feature extraction of the chronic disease sample set to obtain a chronic disease feature set includes: Based on the chi-square test method, feature extraction is performed on the family medical history sample set to obtain a first feature set; Based on the principal component analysis method, feature extraction is performed on the environment sample set and the first lifestyle sample set respectively to obtain a second feature set and a third feature set; Based on the correlation analysis method, feature extraction is performed on the first physiological sample set to obtain a fourth feature set; The first feature set, the second feature set, the third feature set and the fourth feature set are combined in parameter dimension to obtain the chronic disease feature set.

4. The health management system according to claim 3, characterized in that: Initializing a decision tree network architecture according to the training set, and training the decision tree network architecture to obtain a chronic disease training model includes: By using the Bootstrap sampling method, samples are randomly selected from the training set with replacement for multiple times to obtain multiple first training sets and second training sets, wherein the second training set is a sample set that has never been selected from the training set; Initializing multiple decision trees according to the first training sets, and performing split growth on each of the decision trees in a non-pruning strategy to obtain a decision tree network architecture; Inputting the second training set into the decision tree network architecture for multivariate prediction to obtain multiple prediction results; Regression training is performed according to each of the prediction results to obtain a regression prediction architecture, and the regression prediction architecture is merged with the decision tree network architecture to obtain a chronic disease training model; wherein the regression prediction architecture is used to integrate each of the prediction results into an optimal prediction result.

5. The health management system according to claim 4, characterized in that: The process of training and growing the decision tree includes the following steps: Based on the information gain algorithm, the information gain of each sample in the first training set is calculated, and the sample with the maximum information gain is selected as the splitting node; Determine a split value of the split node, and divide the first training set into a plurality of split subsets according to the split value; The multiple split sets are respectively used as the new first training sets, and the information gain algorithm is returned to calculate the information gain of each sample in the first training set, and the sample with the optimal information gain is selected as the split node. This step is repeated until the preset stopping condition is met to obtain the decision tree.

6. The health management system according to claim 1, characterized in that: The cloud server is also used to obtain the physiological data, the lifestyle data and the health change record of the sick member, and input the physiological data, the lifestyle data and the health change record into the health demand library for demand prediction according to the coordination filtering method to obtain the target demand; The cloud server is also used to screen out content consultations that meet the target needs from the health article library and the health service library according to the target needs, and push the content consultations to the home end, wherein the content consultations include health education materials, health activity announcements and health service hotlines.

7. The health management system according to claim 6, characterized in that: The health needs library is obtained by the following steps: Acquire health sample sets and health demand sets of different users, wherein the health sample sets include a second physiological sample set, a second lifestyle sample set, and a health change sample set, and the health demand set includes a plurality of demand samples and demand scores corresponding to the demand samples; Extracting features from each of the healthy sample sets to obtain a plurality of fifth feature sets; Extracting features from each of the health demand sets to obtain a plurality of sixth feature sets; A correlation matrix between the fifth feature set and the sixth feature set of different users is constructed to obtain the health needs library.

8. The health management system according to claim 6, characterized in that: According to the coordinated filtering method, the physiological data, the lifestyle data and the health change record are input into the health demand library for demand prediction to obtain the target demand, including: Calculating a first similarity between each user in the health needs database and the sick member according to the physiological data, the lifestyle data and the health change record, and determining similar users in the health needs database according to the first similarity; Determine the same demand samples in the health demand sets of each of the similar users as similar demands, and determine the demand scores corresponding to each similar demand as similar scores; According to the target similarity between the similar user and the sick member, weighted average calculation is performed on each similar score of the same type to obtain multiple predicted scores; Among the prediction scores, the similar demand corresponding to the prediction score with the highest score is determined as the target demand.

9. The health management system according to claim 8, characterized in that: The calculating, based on the physiological data, the lifestyle data and the health change record, a first similarity between each user in the health needs library and the sick member, and determining similar users in the health needs library based on the first similarity, comprises: Calculating the cosine similarity between the sick member and each user in the health demand database according to the physiological data, the lifestyle data and the health change record, and obtaining a plurality of the first similarities; The first similarity that is greater than or equal to a first preset threshold is determined as a second similarity, and users corresponding to the second similarity in the health needs library are determined as similar users.

10. The health management system according to claim 1, characterized in that: The health management system further includes: A health detection device, the health detection device being communicatively connected to the home terminal; The health detection device is used to collect the physiological data and the environmental data of each family member, and send the physiological data and the environmental data to the family end.

Images