Human health early warning method and system based on physical examination data analysis

Abstract

The invention relates to the technical field of health early warning, in particular to a human health early warning method and system based on physical examination data analysis, which comprises the following steps: collecting physical examination data of a human body, analyzing index changes of blood pressure, blood sugar, cholesterol and body mass index of the human body, sorting the physical examination data of the human body through time sequence sorting, and generating a time sequence sorting result. According to the method, time sequence sorting and spatial dimension evaluation are performed on the physical examination data, so that the health risk can be accurately identified and the early warning level can be dynamically evaluated, the data analysis precision is improved, the risk prediction dimension is increased, early warning is more timely and accurate, and the accuracy of early warning is improved by performing layout evaluation on the physical examination data points in a finite dimension space. Potential health abnormalities can be identified on multiple levels, and multi-dimensional analysis provides a new way for identifying complex health problems, so that health management is more active and predictable, health monitoring and risk management are optimized, and the health condition of a patient is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of health warning, and particularly to a method and system for warning human health based on physical examination data analysis. Background Art

[0002] The technical field of health warning relies on big data analysis, machine learning algorithms, and biometric methods. The aim is to extract useful information from various health monitoring devices and medical records. By continuously monitoring physiological parameters, health indicators, and other relevant data, a health warning system can generate customized alerts to prevent the onset of chronic diseases, acute events, and medical emergencies, playing an important role in improving the quality of life of patients, reducing hospital stays, and optimizing the allocation of medical resources.

[0003] Among them, the method for warning human health based on physical examination data analysis involves using physical examination data to build models and algorithms to predict and warn individuals of health problems. By using various medical data obtained from physical examinations, such as blood analysis results, physiological measurements, and historical health records, through data analysis and pattern recognition techniques, health risks can be identified. Identifying potential health problems in advance enables medical intervention to start before the disease develops to a more severe stage, effectively managing health risks, reducing medical costs, and improving the overall health of patients, providing a powerful prediction tool for medical providers and patients to help make better health management decisions.

[0004] Existing health warning technologies mainly rely on big data analysis and machine learning algorithms for health warning. Although they perform well in processing large amounts of data, they ignore the temporal and spatial dimensional analysis of data, which is particularly crucial when dealing with physical examination data. The efficiency in accurately identifying health trends and real-time dynamic assessment is not high because single-dimensional analysis cannot comprehensively capture all potential health risk factors. For example, for the development of chronic diseases in the early stage, existing health warning technologies fail to provide timely warnings due to the lack of analysis of the layout of data points in multi-dimensional space. Existing technologies also have deficiencies in the customization and personalization of health warnings. Using a general model to process the health data of different individuals and ignoring the differences between individuals lead to inaccurate or delayed warnings, affecting the timeliness and effectiveness of medical intervention. Summary of the Invention

[0005] The purpose of the present invention is to solve the drawbacks existing in the prior art, and to propose a method and system for warning human health based on physical examination data analysis.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions. A method for warning human health based on physical examination data analysis includes the following steps:

[0007] S1: Collect the physical examination data of the human body, analyze the changes in the indicators of blood pressure, blood sugar, cholesterol, and body mass index of the human body, sort out the physical examination data of the human body through time series, and generate a time series sorting result;

[0008] S2: Based on the time series sorting result, evaluate the layout of physical examination data points in a finite-dimensional space, identify potential health abnormal areas, mark the areas, and generate a warning level mapping result;

[0009] S3: Based on the warning level mapping result, analyze the physical examination data points, sort out the data points of the same health condition, and mark them as the target health risk group, and generate a health risk group identification result;

[0010] S4: Utilize the health risk group identification result, analyze the physical examination data points of the risk group, capture the relationships between health indicators, identify key risk factors, and generate associated health risk factors;

[0011] S5: Based on the associated health risk factors, conduct a time series analysis of the human health warning data, predict the changes in the human health status in the future time period, identify potential health trends, and generate a health trend prediction result;

[0012] S6: Adopt the health trend prediction result, monitor the key health indicators that deviate from the standard threshold, dynamically evaluate the warning level of human health, analyze potential health risks, and generate a health risk analysis result.

[0013] As a further solution of the present invention, the time series sorting result includes rearranged data points, timestamps of the data points, and health indicator values of the data points. The warning level mapping result includes abnormal thresholds of health indicators, geographical mapping of marked health abnormal areas, and warning levels of marked areas. The health risk group identification result includes identified health risk levels, health status labels of the group, and the number and characteristics of the risk group. The associated health risk factors include identified key health risk factors, correlation scores between risk factors, and a list of key health indicators. The health trend prediction result includes predicted trends of changes in health indicators, potential risk points, and predicted risk occurrence times. The health risk analysis result includes a list of health indicators exceeding the standard threshold, dynamically evaluated warning levels, and potential health risks.

[0014] As a further solution of the present invention, the steps of collecting the physical examination data of the human body, analyzing the changes in the indicators of blood pressure, blood sugar, cholesterol, and body mass index of the human body, and sorting out the physical examination data of the human body through time series to generate a time series sorting result are specifically as follows:

[0015] S101: Collect the physical examination data of the human body, calibrate the physical examination time through the time stamp, match the physical examination date for the physical examination data record, check the consistency of the time line, and generate the time stamp calibration result;

[0016] S102: Based on the time stamp calibration result, analyze the seasonal change patterns of blood pressure, blood sugar, cholesterol, and body mass index, identify the fluctuations of health indicators over time, and generate the seasonal analysis result of health indicators;

[0017] S103: According to the seasonal analysis result of the health indicators, perform time series sorting on the physical examination data, arrange the physical examination data in order according to the physical examination date, and generate the time series sorting result.

[0018] As a further solution of the present invention, based on the time series sorting result, the steps of evaluating the layout of physical examination data points in a finite-dimensional space, identifying potential health abnormal areas, marking the areas, and generating the warning level mapping result are specifically as follows:

[0019] S201: Based on the time series sorting result, perform position evaluation on the physical examination data points in a finite-dimensional space, track and locate the real-time changes of the physical examination data points, verify the layout of the physical examination data, and generate the spatial layout evaluation result;

[0020] S202: According to the spatial layout evaluation result, detect the abnormal areas of the physical examination data points, identify the data areas that deviate from the conventional health indicators, and identify the deviation areas by setting the thresholds of the health indicators, and generate the identification result of potential health abnormal areas;

[0021] S203: Based on the identification result of the potential health abnormal areas, assign warning levels to the marked areas, and classify the warning levels according to the urgency of the potential health risks, and generate the warning level mapping result.

[0022] As a further solution of the present invention, based on the warning level mapping result, the steps of analyzing the physical examination data points, sorting out the data points of the same health condition, and marking them as the target health risk group, and generating the health risk group identification result are specifically as follows:

[0023] S301: Based on the warning level mapping result, initialize the grouping of the physical examination data points according to the preset risk levels, summarize the physical examination data points and the associated risk categories, and generate the health risk grouping result;

[0024] S302: According to the health risk grouping result, adopt the decision tree algorithm to perform multi-dimensional health index analysis on the physical examination data of the human body, and perform iterative classification according to the index fitting degree, and generate the health condition clustering result;

[0025] S303: Using the clustering results of the health status, mark the health risk levels for the datasets of multiple clusters, verify that the risk levels of the group are consistent with the warning levels, and generate the identification results of the health risk groups.

[0026] As a further solution of the present invention, the formula of the decision tree algorithm is as follows:

[0027]

[0028] Wherein, G(T) is the total impurity of node T, k is the number of child nodes, n i is the number of samples in the i-th child node, n is the total number of samples in the parent node T, and H(T i ) is the impurity of the i-th child node, λ, ρ, and σ are weight coefficients, Var, Entropy, and Skew are respectively used to control variance, entropy, and skewness, |T i | and |T| are the sample weights of the child node and the parent node respectively.

[0029] As a further solution of the present invention, using the identification results of the health risk groups, analyzing the physical examination data points of the risk groups, capturing the relationships between health indicators, identifying key risk factors, and generating the steps of associated health risk factors are specifically as follows:

[0030] S401: Based on the identification results of the health risk groups, screen the physical examination data points of the risk groups, centrally analyze the target health indicators of the risk groups, analyze the changes in blood pressure and blood sugar, and generate the results of the screened risk group data points;

[0031] S402: Analyze the results of the screened risk group data points, compare the correlations of health indicators, analyze the correlation relationships and effects between different health indicators, including blood pressure and cholesterol, and generate the health indicator correlation analysis results;

[0032] S403: According to the health indicator correlation analysis results, identify the key health indicators associated with the risk status, mark them as key risk factors, evaluate the degree of influence of health indicators on health risks, and generate associated health risk factors.

[0033] As a further solution of the present invention, based on the associated health risk factors, perform time series analysis on the human health warning data, predict the changes in the health status of the human body in the future time period, identify potential health trends, and generate the steps of the health trend prediction results are specifically as follows:

[0034] S501: Collect human health warning data associated with health risk factors according to the said associated health risk factors, analyze the change patterns of human heart rate, blood pressure and sleep quality, perform time marking, verify the time series of human health data, and generate a marked health time series data set;

[0035] S502: Use the said marked health time series data set to analyze the changes of human health indicators over time, extract the change trends, including heart rate rate and blood pressure fluctuation frequency, analyze the time records of human medication and exercise, identify the key time nodes affecting health, and obtain the analysis results of key time points;

[0036] S503: According to the analysis results of the said key time points, identify the periodicity and development direction of health status, analyze the changes of disease risks and health indicators, predict the potential health risks and risk trends within a future time period, and generate a health trend prediction result.

[0037] As a further solution of the present invention, the steps of using the said health trend prediction result to monitor key health indicators deviating from the standard threshold, dynamically evaluate the warning level of human health, analyze potential health risks, and generate a health risk analysis result are specifically as follows:

[0038] S601: Based on the said health trend prediction result, set monitoring parameters, track key health indicators in real time, analyze the periodicity of heart rate and cholesterol levels, and identify data points deviating from the normal range in real time to generate a monitoring result of key health indicators;

[0039] S602: According to the monitoring result of the said key health indicators, conduct trend analysis and deviation evaluation on the monitored health indicators, analyze the warning signals of human health in combination with the patient's historical health records and lifestyle information, and generate a dynamic health assessment result;

[0040] S603: Use the said dynamic health assessment result to analyze the association between the recorded health indicator data and potential health risks, identify risk factors, evaluate the development trend and impact of potential health risks, and generate a health risk analysis result

[0041] A human health warning system based on physical examination data analysis, the human health warning system based on physical examination data analysis is used to execute the above-mentioned human health warning method based on physical examination data analysis, and the system includes:

[0042] The time sorting module sorts the physical examination data points based on the historical physical examination data of human blood pressure, blood sugar, cholesterol and body mass index, and generates a time series sorting result;

[0043] Based on the time series sorting result, the spatial layout evaluation module analyzes the distribution of physical examination data points in a finite-dimensional space, identifies areas of health anomalies, marks them, and generates a warning level mapping result;

[0044] Based on the warning level mapping result, the risk clustering module performs clustering analysis on the health indicators of physical examination data points, identifies and marks groups with similar health risk characteristics, and generates a health risk group identification result;

[0045] Based on the health risk group identification result, the risk factor association module analyzes the physical examination data in different groups, captures the correlation between health indicators, and identifies key health risk factors by comparing the differences and relationships of health indicators, generating associated health risk factors;

[0046] Based on the associated health risk factors, the health trend prediction module analyzes health warning signals in the future time period, identifies potential health trends, and analyzes the changes and impacts of health risks, generating a health risk analysis result.

[0047] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0048] In the present invention, by collecting and analyzing physical examination data, performing time series sorting and spatial dimension evaluation on the data, health risks can be accurately identified and the warning level can be dynamically evaluated. This not only improves the accuracy of data analysis but also increases the dimension of risk prediction, making the warning more timely and accurate. Through the layout evaluation of physical examination data points in a finite-dimensional space, potential health anomalies can be identified at multiple levels, and multi-dimensional analysis provides a new way to identify complex health problems. By dynamically linking the correlations between health indicators, the ability to interpret health risk factors is strengthened, making health management more proactive and predictive, optimizing health monitoring and risk management, reducing medical costs, and improving the health status of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a schematic diagram of the working process of the present invention;

[0050] Figure 2 is a detailed flowchart of S1 of the present invention;

[0051] Figure 3 is a detailed flowchart of S2 of the present invention;

[0052] Figure 4 is a detailed flowchart of S3 of the present invention;

[0053] Figure 5 is a detailed flowchart of S4 of the present invention;

[0054] Figure 6The refined flowchart of S5 of the present invention;

[0055] Figure 7 The refined flowchart of S6 of the present invention;

[0056] Figure 8 The system flowchart of the present invention. Detailed implementation manners

[0057] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0058] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention 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 of the present invention. In addition, in the description of the present invention, "a plurality of" means two or more unless otherwise specifically defined.

[0059] Embodiment 1

[0060] Please refer to Figure 1 , the present invention provides a technical solution, a method for predicting human health based on physical examination data analysis, including the following steps:

[0061] S1: Collect the physical examination data of the human body, analyze the index changes of the human body's blood pressure, blood sugar, cholesterol and body mass index, sort and organize the physical examination data of the human body through time series, identify the fluctuations of health indicators over time, and generate a time series sorting result;

[0062] S2: Based on the time series sorting result, evaluate the layout of physical examination data points in a finite-dimensional space, identify potential health abnormal areas, set the threshold of health indicators to identify deviation areas, and generate a warning level mapping result;

[0063] S3: Based on the warning level mapping result, analyze the physical examination data points, sort out the data points of the same health status, and mark them as the target health risk group, and generate a health risk group identification result;

[0064] S4: Use the health risk group identification result to analyze the physical examination data points of the risk group, compare the correlation of health indicators, capture the relationship between health indicators, identify key risk factors, and generate associated health risk factors;

[0065] S5: Based on the associated health risk factors, conduct time - series analysis of the human health warning data, predict the changes in the human health status within a future time period, identify potential health trends, and generate health trend prediction results;

[0066] S6: Use the health trend prediction results to monitor the key health indicators that deviate from the standard thresholds, dynamically evaluate the warning level of human health, analyze potential health risks, and generate health risk analysis results.

[0067] The time - series sorting results include the rearranged data points, the timestamps of the data points, and the health indicator values of the data points. The warning level mapping results include the abnormal thresholds of the health indicators, the geographical mapping of the marked health abnormal areas, and the warning levels of the marked areas. The health risk group identification results include the identified health risk levels, the health status labels of the groups, and the number and characteristics of the risk groups. The associated health risk factors include the identified key health risk factors, the correlation scores between the risk factors, and the list of key health indicators. The health trend prediction results include the predicted change trends of the health indicators, the potential risk points, and the predicted risk occurrence times. The health risk analysis results include the list of health indicators exceeding the standard thresholds, the dynamically evaluated warning levels, and the potential health risks.

[0068] Please refer to Figure 2 , collect the physical examination data of the human body, analyze the changes in the indicators of human blood pressure, blood sugar, cholesterol, and body mass index, and sort the physical examination data of the human body through time - series sorting. The specific steps for generating the time - series sorting results are as follows:

[0069] S101: Collect the physical examination data of the human body, calibrate the physical examination time through the timestamp, match the physical examination date for the physical examination data record, check the consistency of the time line, and the execution process for generating the timestamp calibration result is as follows;

[0070] The sub - steps of S101 collect the physical examination data of the human body. For each physical examination data, including parameters such as heart rate, blood pressure, blood sugar, and cholesterol, record the corresponding timestamp information. Each piece of data will be initially recorded according to the original timestamp recorded at the physical examination time, and the timestamp will be calibrated to eliminate the problem of device time asynchronization or time - marking errors. The specific method is to set a reference timestamp, compare the timestamp of each physical examination data, and calculate the difference from the reference timestamp. According to the difference, adjust each timestamp to accurately match the physical examination date, ensure the consistency of the time line of all physical examination records, and generate the timestamp calibration result. The formula used is:

[0071]

[0072] where, T′ i represents the calibrated timestamp, and Tb is the reference timestamp, n is the total number of data samples, and ΔT i is the difference between the original timestamp and the reference timestamp.

[0073] S102: Based on the timestamp calibration result, analyze the seasonal change patterns of blood pressure, blood sugar, cholesterol, and body mass index, identify the fluctuations of health indicators over time, and the execution process for generating the seasonal analysis result of health indicators is as follows;

[0074] The sub-step of S102, based on the timestamp calibration result, collects the data of blood pressure, blood sugar, cholesterol, and body mass index corresponding to each season. By collecting data throughout the four seasons, a seasonal analysis model is used to identify the patterns of fluctuations of health indicators over time. The model analysis includes calculating the average health indicator value for each season and comparing the health change trends in different seasons. By statistically analyzing the average values of each indicator in spring, summer, autumn, and winter, health risks and preventive measures are identified. The analysis result helps to understand how health indicators are affected by seasonal changes and generates the seasonal analysis result of health indicators. The formula used is:

[0075]

[0076] where, H avg,s is the average health indicator for season s, H i is the health indicator value of the i-th data point in season s, and n s is the total number of data points in season s.

[0077] S103: According to the seasonal analysis result of health indicators, perform a time series sorting on the physical examination data, arrange the physical examination data in order according to the physical examination date, and the execution process for generating the time series sorting result is as follows;

[0078] The sub-step of S103, according to the seasonal analysis result of health indicators, performs a time series sorting on the physical examination data, classifies the physical examination data according to the physical examination date, and each classification represents a specific month or season. Inside each classification, the physical examination data is arranged in the order of the physical examination date. This step ensures that the time order of the data is consistent with the actual physical examination time order. The sorting of the data not only reflects the sequence of time but also maintains the internal consistency and comparability of the data, generating the time series sorting result. The formula used is:

[0079] R sorted = sort(D, key = T)

[0080] where, R sorted is the sorted time series result, D is the physical examination data set, T is the timestamp field of the physical examination data, and sort is the function for sorting by timestamp.

[0081] Please refer to Figure 3, based on the time series sorting results, evaluate the layout of physical examination data points in a finite-dimensional space, identify potential health abnormal areas, mark the areas, and the specific steps for generating the warning level mapping results are as follows:

[0082] S201: Based on the time series sorting results, evaluate the positions of physical examination data points in a finite-dimensional space, track and locate the real-time changes of physical examination data points, verify the layout of physical examination data, and the execution process for generating the spatial layout evaluation results is as follows;

[0083] The sub-step of S201 maps physical examination data points into a finite-dimensional space based on the time series sorting results. Each dimension represents a health indicator such as blood pressure or cholesterol level. By constructing a multi-dimensional data model, the position of each physical examination data point in the model represents the specific value on each health indicator. Track the real-time changes of the data points in the subsequent time period, such as the increase or decrease of health indicators, to verify whether the relative position relationship between the data points conforms to the expected health change trend. This can not only check the status of a single data point but also observe the overall layout of the data group, and generate the spatial layout evaluation results.

[0084] S202: According to the spatial layout evaluation results, detect abnormal areas of physical examination data points, identify data areas that deviate from the conventional health indicators, and mark the deviation areas by setting the threshold of health indicators. The execution process for generating the potential health abnormal area identification results is as follows;

[0085] The sub-step of S202 detects abnormal areas of physical examination data points according to the spatial layout evaluation results, sets the normal range for each health indicator, and identifies abnormalities by comparing whether the health indicators of each data point exceed the range. For data points that exceed the normal range, they are marked as deviation areas. The deviation areas indicate potential health problems. By calculating the density of data points in each deviation area, further identify the areas with a high density of abnormal data points, and generate the potential health abnormal area identification results. The formula used is:

[0086]

[0087] where D r represents the density of abnormal data points in a certain area, and 1 outofrange (P i ) is an indicator function, which takes the value of 1 when the data point P i exceeds the normal range, and 0 otherwise. n is the total number of data points in the area.

[0088] S203: Based on the potential health abnormal area identification results, assign warning levels to the marked areas. The warning levels are classified according to the urgency of potential health risks. The execution process for generating the warning level mapping results is as follows;

[0089] Based on the recognition results of potential health abnormal regions, the sub-step S203 assigns a warning level to each marked region. The warning levels are classified according to the urgency of potential health risks, such as low, medium, and high. A grading method is used to jointly determine the warning level based on the data point density and the deviation degree of health indicators in the abnormal region. Regions with high density or large deviation degree obtain a higher warning level, which not only quantifies the health risks but also provides a basis for taking different preventive measures for different levels of risks, generating a warning level mapping result.

[0090] Please refer to Figure 4 , based on the warning level mapping result, the steps of analyzing the physical examination data points, sorting out the data points with the same health status, and marking them as the target health risk group to generate the health risk group recognition result are as follows:

[0091] S301: Based on the warning level mapping result, the physical examination data points are initially grouped according to the preset risk levels, and the physical examination data points and the associated risk categories are summarized. The execution process of generating the health risk grouping result is as follows;

[0092] The sub-step S301 of S301 initially groups the physical examination data points according to the preset risk levels based on the warning level mapping result, compares each index of each data point with the corresponding rules of the risk levels, such as comparing indexes such as blood pressure and heart rate with the risk thresholds. According to the comparison results, the physical examination data points are initially grouped. For each physical examination data point, the number of conditions that meet each risk level is calculated, and the risk level with the largest number of conditions is used as the initial risk level of the data point. To accurately map the risk levels, for the data points in the boundary situation, the deviation value of the critical index is calculated to determine the risk level, and the data points are summarized according to the risk levels to generate the health risk grouping result. The formula is:

[0093]

[0094] Among them, H r represents the health risk score, n is the number of risk indicators, w i is the weight of the i-th index, and δ is a function that takes 1 when the physical examination index p i exceeds or is equal to the risk threshold t i and takes 0 otherwise.

[0095] S302: According to the health risk grouping result, using the decision tree algorithm, the multi-dimensional health indicators of the physical examination data of the human body are analyzed, and iterative classification is carried out according to the index fitting degree. The execution process of generating the health status clustering result is as follows;

[0096] In sub-step S302, based on the health risk grouping results, multi-dimensional health index analysis is performed on the physical examination data of the human body. The physical examination data is sorted and classified according to the initial risk grouping results, and each category contains physical examination data with similar risk levels. Detailed analysis is carried out on the data of each category, and the median, average value, and standard deviation of each health index are calculated respectively, as well as the distribution differences of the index among different risk levels. Through the analysis of the statistical data, the contribution degree of each health index in the risk classification is evaluated, and the weight allocation of the index is carried out accordingly. The data points are reclassified according to the index values adjusted by the weights, and the classification accuracy is iteratively optimized until the classification results are stable, generating the health status clustering results.

[0097] The formula of the decision tree algorithm is as follows:

[0098]

[0099] Among them, G(T) is the total impurity of node T, k is the number of child nodes, n i is the number of samples in the i-th child node, n is the total number of samples in the parent node T, H(T i ) is the impurity of the i-th child node, λ, ρ, and σ are weight coefficients, Var, Entropy, and Skew are used to control variance, entropy, and skewness respectively, |T i | and |T| are the sample weights of the child node and the parent node respectively.

[0100] The execution process is as follows:

[0101] Calculate the impurity H(T i ) of each child node and the proportion of the impurity in the parent node, add the proportion of the number of samples in each child node to the parent node, and increase the adjustment terms of variance Var(T i ), entropy Entropy(T i ), and skewness Skew(T i ). The statistical metric provides additional discriminative power for the construction of the decision tree. The splitting effect of the model is optimized by determining the values of the weights λ, ρ, and σ. The determination of the weights can be selected through cross-validation or an optimization algorithm based on prior knowledge. The entire calculation process is iterated until the set tree depth is reached or specific stopping conditions are met.

[0102] S303: Using the health status clustering results, label the health risk levels for the data sets of multiple clusters, verify that the risk levels of the group are consistent with the warning levels, and the execution process of generating the health risk group identification results is as follows;

[0103] In sub-step S303, the health risk levels are labeled for the datasets of multiple clusters by using the clustering results of health conditions. The data points in each clustering result are re-evaluated for risk levels according to their positions in the clusters and the performance of health indicators. By comparing the clustering results with the preset risk thresholds, it is verified whether the health conditions of the data points in each cluster are consistent with the warning levels. If the clustering indicators of the data points do not match the warning levels, the risk levels are adjusted until the average risk scores of each cluster meet the requirements of the warning levels. Through statistical analysis, the identification results of health risk groups are generated.

[0104] Please refer to Figure 5 , by using the identification results of health risk groups, analyze the physical examination data points of the risk groups, capture the relationships between health indicators, identify key risk factors, and the steps for generating associated health risk factors are specifically as follows:

[0105] S401: Based on the identification results of health risk groups, screen the physical examination data points of the risk groups, centrally analyze the target health indicators of the risk groups, analyze the changes in blood pressure and blood sugar, and the execution process for generating the screened results of the risk group data points is as follows;

[0106] In sub-step S401, based on the identification results of health risk groups, screen the physical examination data points of the risk groups. First, extract all the physical examination data points classified as high-risk from the identification results of health risk groups. Further analyze the data points, focusing on the target health indicators such as blood pressure and blood sugar. Conduct a detailed statistics on the blood pressure and blood sugar records of each data point, calculate the average value, fluctuation range, and change trend. According to the time series of the data points, evaluate the change rules of blood pressure and blood sugar over time, especially the changes before and after the physical examination. Summarize the analysis results to generate the screened results of the risk group data points, and the formula used is:

[0107]

[0108] where R d represents the risk measure of the health indicator, n is the number of data points, d k is the blood pressure or blood sugar value of the kth data point, is the average blood pressure or blood sugar value of the data points.

[0109] S402: Analyze the screened results of the risk group data points, compare the correlations of the health indicators, and analyze the correlation relationships and effects between the differential health indicators, including blood pressure and cholesterol, and the execution process for generating the health indicator correlation analysis results is as follows;

[0110] The sub-step of S402 analyzes the results of the screened risk group data points, compares the correlations of health indicators, and analyzes the correlations between indicators such as blood pressure and cholesterol through statistical methods. Using correlation coefficient calculation, determine the strength of the relationship between blood pressure and cholesterol, and further reveal the specific association patterns between indicators through scatter plots and regression analysis. Mark the data points of outliers specially, and analyze the reasons for the outliers and special cases of their impact on health. Integrate the analysis results to generate the analysis results of health indicator associations. The formula used is:

[0111]

[0112] Among them, C xy represents the correlation coefficient between blood pressure x and cholesterol y, x i and y i are the blood pressure and cholesterol values of a single data point, and are the average values of the values.

[0113] S403: According to the analysis results of health indicator associations, identify the key health indicators associated with the risk status, mark them as key risk factors, and evaluate the degree of influence of health indicators on health risks. The execution process for generating associated health risk factors is as follows;

[0114] The sub-step of S403 identifies the key health indicators associated with the risk status according to the analysis results of health indicator associations, and selects the indicators highly correlated with health risks as key risk factors. By comparing the distributions and statistical characteristics of key indicators in different risk-level groups, evaluate the degree of influence of key indicators on health risks. Use multiple regression analysis to calculate the weights and contribution rates of each key indicator in health risk assessment, and more precisely quantify the impact on overall health risks to generate associated health risk factors. The formula used is:

[0115] F r =β0 + β1x1 + β2x2 + … + β n x n

[0116] Among them, F r represents the health risk score, β0 is the constant term, and β n is the regression coefficient of the nth key health indicator x n of.

[0117] Please refer to Figure 6 , based on the associated health risk factors, conduct time series analysis of human health warning data, predict the changes in human health status in the future time period, identify potential health trends, and the steps for generating health trend prediction results are specifically:

[0118] S501: According to the associated health risk factors, collect human health warning data associated with the health risk factors, analyze the change patterns of human heart rate, blood pressure and sleep quality, and perform time marking to verify the timeliness of human health data. The execution process of generating the marked health time series dataset is as follows;

[0119] The sub - steps of S501 collect human health warning data associated with the health risk factors according to the associated health risk factors, determine the health indicators closely related to the key health risk factors, such as heart rate, blood pressure and sleep quality, and extract the data of the indicators from the historical medical records. Perform time marking on each data point to ensure the accuracy of the time information of each record. Analyze the daily change patterns of heart rate, blood pressure and sleep quality, use time - series analysis techniques to observe the performance of the indicators at different time periods (such as morning, evening, after exercise), verify the timeliness of human health data, and process any abnormal data points to ensure the accuracy and integrity of the dataset, generating the marked health time series dataset. The formula is:

[0120]

[0121] where, T s represents the health indicator score of the time series, α is the adjustment coefficient, n is the number of data points, h i is the health indicator value of the i - th data point, and t i is the time weight of the data point.

[0122] S502: Use the marked health time series dataset to analyze the changes of human health indicators over time, extract the change trends, including the heart rate rate and blood pressure fluctuation frequency, analyze the time records of human medication and exercise, identify the key time nodes affecting health, and obtain the execution process of the key time point analysis result as follows;

[0123] The sub - steps of S502 use the marked health time series dataset to analyze the changes of human health indicators over time, evaluate the stability and change trends of the indicators over time by calculating the change rate of heart rate and the fluctuation frequency of blood pressure. Analyze the external activity data including the time records of medication and exercise, identify the key time nodes where activities affect health indicators. Compare the performance of health indicators in different time periods, use statistical methods to identify the time points significantly related to health changes, and generate the key time point analysis result. The formula is:

[0124]

[0125] where, V t represents the health indicator change trend value, n is the number of key time points, m is the number of health indicators, β j is the influence coefficient of indicator j, f(xj ,t j ) is the function expression of the index x j at time t j .

[0126] S503: According to the analysis results of key time points, identify the periodicity and development direction of the health status, analyze the changes in disease risks and health indicators, predict potential health risks and risk trends within a future time period, and the execution process of generating the health trend prediction results is as follows;

[0127] The sub-steps of S503 identify the periodicity and development direction of the health status according to the analysis results of key time points. Analyze the relationship between disease risks and health indicators over time, and identify regular changes and abnormal patterns in the health status through periodic analysis. Use the information to predict potential health risks and risk trends within a future time period, and comprehensively utilize historical data and current analysis results to estimate the future health risk probability, generating the health trend prediction results. The formula used is:

[0128]

[0129] Among them, P h represents the health risk prediction score, γ is the standardization coefficient, q is the number of periodic indicators, θ k is the importance weight of the index y k . g(y k ) is the health risk function of the index y k .

[0130] Please refer to Figure 7 , adopt the health trend prediction results, monitor key health indicators that deviate from the standard thresholds, dynamically evaluate the early warning level of human health, analyze potential health risks, and the steps of generating the health risk analysis results are specifically as follows:

[0131] S601: Based on the health trend prediction results, set monitoring parameters, track key health indicators in real time, analyze the periodicity of heart rate and cholesterol levels, and identify data points that deviate from the normal range in real time. The execution process of generating the key health indicator monitoring results is as follows;

[0132] Sub-step S601 sets monitoring parameters based on the health trend prediction results, tracks key health indicators in real time, determines the key health indicators that need to be continuously monitored, including heart rate and cholesterol level, and sets the normal range of the indicators according to the previous health trend prediction. Use real-time data collection devices such as wearable health monitors to collect and upload heart rate and cholesterol data in real time. Analyze the periodicity and abnormal fluctuations of the data in real time through the set algorithm, and identify any data points that exceed the normal range. For the detected abnormal data, further analysis is carried out to confirm whether it is a real health risk or just a data collection error, and generate the monitoring results of key health indicators. The formula is:

[0133]

[0134] where M k represents the monitoring score of the key health indicator, N is the number of data points, x i is the health indicator value of the i-th data point, μ and σ are the mean and standard deviation of the indicator respectively, and σ is the adjustment function.

[0135] S602: According to the monitoring results of key health indicators, conduct trend analysis and deviation assessment on the monitored health indicators, combine the patient's historical health records and lifestyle information, analyze the warning signals of human health, and the execution process of generating dynamic health assessment results is as follows;

[0136] Sub-step S602 conducts trend analysis and deviation assessment on the monitored health indicators according to the monitoring results of key health indicators. For each monitored health indicator, such as heart rate, blood pressure, etc., time series analysis is carried out to identify the change trend and periodic fluctuations. Combine the patient's historical health records and current lifestyle information, such as eating habits, sleep patterns, and physical activity levels, and use multiple regression analysis to evaluate the specific impact of lifestyle factors on health indicators. Warning signals related to health changes can be identified, and the impact of individual differences and environmental factors on health status is considered. For all abnormal or significantly deviated health indicator data points, detailed analysis is carried out to determine potential health risks or non-disease-related variation reasons, and generate dynamic health assessment results. The formula is:

[0137]

[0138] where A t represents the dynamic health assessment score, λ0 is the constant term, λ i is the coefficient of health indicator h i , μ j is the coefficient of lifestyle indicator l j , and p and q are the numbers of health indicators and lifestyle indicators respectively.

[0139] S603: The execution process of using the dynamic health assessment results to analyze the association between the recorded health indicator data and potential health risks, identify risk factors, evaluate the development trend and impact of potential health risks, and generate health risk analysis results is as follows;

[0140] The sub-steps of S603 use the dynamic health assessment results to analyze the association between the recorded health indicator data and potential health risks, identify risk factors, and through an integrated analysis method, compare the health indicator data with known disease risk factors such as heart disease, diabetes, etc., to evaluate the statistical correlation between health indicators and specific health risks. Analyze the change pattern of health indicators over time through a machine learning model to predict the development trend of potential health risks. Cover the identification of risk factors, evaluate the degree of impact on individual health, and consider relevant factors such as genetic factors and environmental stress to generate health risk analysis results. The formula used is:

[0141]

[0142] Where, R f represents the health risk analysis score, η0 is the baseline risk value, and η k is the weight of the factor f k related to a specific health risk, and r is the number of risk factors.

[0143] Please refer to Figure 8 , a human health warning system based on physical examination data analysis. The human health warning system based on physical examination data analysis is used to execute the above-mentioned human health warning method based on physical examination data analysis. The system includes:

[0144] The time sorting module sorts the physical examination data points based on the historical physical examination data of human blood pressure, blood sugar, cholesterol, and body mass index, and generates a time series sorting result;

[0145] The spatial layout evaluation module analyzes the distribution of physical examination data points in a finite-dimensional space based on the time series sorting result, identifies the areas of health abnormalities, and marks them to generate a warning level mapping result;

[0146] The risk clustering module performs cluster analysis on the health indicators of physical examination data points based on the warning level mapping result, identifies and marks the groups with similar health risk characteristics, and generates a health risk group identification result;

[0147] The risk factor association module analyzes the physical examination data in the differentiated groups based on the health risk group identification result, captures the association between health indicators, and identifies the key health risk factors by comparing the differences and relationships of health indicators to generate associated health risk factors;

[0148] Based on associated health risk factors, the health trend prediction module analyzes health warning signals within a future time period, identifies potential health trends, and analyzes the changes and impacts of health risks to generate health risk analysis results.

[0149] The above are only the preferred embodiments of the present invention, and there are no other forms of limitation to the present invention. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for human health warning based on physical examination data analysis, characterized in that It includes the following steps: Collect the physical examination data of the human body, analyze the index changes of the human body's blood pressure, blood sugar, cholesterol and body mass index, sort out the physical examination data of the human body through time series, and generate a time series sorting result; Based on the time series sorting result, evaluate the layout of physical examination data points in a finite-dimensional space, identify potential health abnormal areas, mark the areas, and generate a warning level mapping result; Based on the warning level mapping result, analyze the physical examination data points, sort out the data points of the same health condition, and mark them as the target health risk group, and generate a health risk group identification result; Using the health risk group identification result, analyze the physical examination data points of the risk group, capture the relationships between health indicators, identify key risk factors, and generate associated health risk factors; Based on the associated health risk factors, conduct a time series analysis of the human health warning data, predict the changes in the human health status in the future time period, identify potential health trends, and generate a health trend prediction result; Adopt the health trend prediction result, monitor the key health indicators that deviate from the standard threshold, dynamically evaluate the warning level of human health, analyze potential health risks, and generate a health risk analysis result.

2. The method for human health warning based on physical examination data according to claim 1, wherein, The time series sorting result includes rearranged data points, timestamps of data points, and health indicator values of data points. The warning level mapping result includes abnormal thresholds of health indicators, geographical mapping of marked health abnormal areas, and warning levels of marked areas. The health risk group identification result includes identified health risk levels, health status labels of groups, and the quantity and characteristics of risk groups. The associated health risk factors include identified key health risk factors, correlation scores between risk factors, and a list of key health indicators. The health trend prediction result includes predicted health indicator change trends, potential risk points, and predicted risk occurrence times. The health risk analysis result includes a list of health indicators exceeding the standard threshold, dynamically evaluated warning levels, and potential health risks.

3. The method for warning of human body health based on physical examination data according to claim 1, characterized in that, The steps of collecting the physical examination data of the human body, analyzing the index changes of the human body's blood pressure, blood sugar, cholesterol and body mass index, and sorting out the physical examination data of the human body through time series to generate a time series sorting result are specifically as follows: Collect the physical examination data of the human body, calibrate the physical examination time through timestamps, match the physical examination date for the physical examination data record, check the consistency of the time line, and generate a timestamp calibration result; Based on the timestamp calibration result, analyze the seasonal change patterns of blood pressure, blood sugar, cholesterol and body mass index, identify the fluctuations of health indicators over time, and generate a health indicator seasonal analysis result; According to the health indicator seasonal analysis result, sort the physical examination data in time series, arrange the physical examination data in order according to the physical examination date, and generate a time series sorting result.

4. The method for predicting human health warnings based on physical examination data according to claim 1, wherein, The steps of evaluating the layout of physical examination data points in a finite-dimensional space, identifying potential health abnormal areas, marking the areas, and generating a warning level mapping result based on the time series sorting result are specifically as follows: Based on the time series sorting result, evaluate the position of the physical examination data points in a finite-dimensional space, track and locate the real-time changes of the physical examination data points, verify the layout of the physical examination data, and generate a spatial layout evaluation result; According to the spatial layout evaluation result, detect the abnormal areas of the physical examination data points, identify the data areas that deviate from the conventional health indicators, mark the deviation areas by setting the threshold of the health indicators, and generate a potential health abnormality area identification result; Based on the potential health abnormality area identification result, assign a warning level to the marked area. The warning level is classified according to the urgency of the potential health risk, and a warning level mapping result is generated.

5. The method for predicting human health warnings based on physical examination data according to claim 1, characterized in that, Based on the warning level mapping result, the steps of analyzing the physical examination data points, sorting out the data points of the same health condition, and marking them as the target health risk group to generate a health risk group identification result are specifically as follows: Based on the warning level mapping result, initialize the grouping of the physical examination data points according to the preset risk levels, summarize the physical examination data points and the associated risk categories, and generate a health risk grouping result; According to the health risk grouping result, use the decision tree algorithm to perform multi-dimensional health index analysis on the physical examination data of the human body, and perform iterative classification according to the index fitting degree to generate a health condition clustering result; Using the health condition clustering result, mark the health risk levels for the data sets of multiple clusters, verify that the risk levels of the groups are consistent with the warning levels, and generate a health risk group identification result.

6. The method for human health warning based on physical examination data according to claim 5, wherein, The formula of the decision tree algorithm is as follows: Among them, G(T) is the total impurity of node T, k is the number of child nodes, n i is the number of samples in the i-th child node, n is the total number of samples in the parent node T, H(T i ) is the impurity of the i-th child node, λ, ρ, and σ are weight coefficients, Var, Entropy, and Skew are used to control variance, entropy, and skewness respectively, |T i | and |T| are the sample weights of the child node and the parent node respectively.

7. The method for warning of human body health based on physical examination data according to claim 1, wherein Using the health risk group identification result, analyze the physical examination data points of the risk group, capture the relationships between the health indicators, and identify the key risk factors. The steps of generating the associated health risk factors are specifically as follows: Based on the health risk group identification result, screen the physical examination data points of the risk group, centrally analyze the target health indicators of the risk group, and analyze the changes in blood pressure and blood sugar to generate a result of the screened risk group data points; Analyze the result of the screened risk group data points, compare the relevance of the health indicators, and analyze the association relationship and role between the differential health indicators, including blood pressure and cholesterol, to generate a health indicator association analysis result; According to the health indicator association analysis result, identify the key health indicators associated with the risk status, mark them as key risk factors, and evaluate the degree of influence of the health indicators on the health risk to generate associated health risk factors.

8. The method for human health warning based on physical examination data analysis according to claim 1, characterized in that, Based on the associated health risk factors, the steps of performing time series analysis on the human health warning data, predicting the changes in the human health status in the future time period, and identifying the potential health trends to generate a health trend prediction result are specifically as follows: According to the associated health risk factors, collect the human health warning data associated with the health risk factors, analyze the change patterns of the human heart rate, blood pressure and sleep quality, and perform time marking to verify the timeliness of the human health data, and generate a marked health time series data set; Using the marked healthy time series dataset, analyze the changes of human health indicators over time, extract the change trends, including heart rate rate and blood pressure fluctuation frequency, analyze the time records of human medication and exercise, identify the key time nodes affecting health, and obtain the analysis results of key time points; According to the analysis results of the key time points, identify the periodicity and development direction of the health status, analyze the disease risk and the changes of health indicators, predict the potential health risks and risk trends in the future time period, and generate the health trend prediction results.

9. The method for warning of human body health based on physical examination data according to claim 1, characterized in that, The steps of using the health trend prediction results to monitor the key health indicators deviating from the standard threshold, dynamically evaluate the warning level of human health, analyze the potential health risks, and generate the health risk analysis results are specifically as follows: Based on the health trend prediction results, set the monitoring parameters, track the key health indicators in real time, analyze the periodicity of heart rate and cholesterol levels, and identify the data points deviating from the normal range in real time to generate the monitoring results of key health indicators; According to the monitoring results of the key health indicators, conduct trend analysis and deviation evaluation on the monitored health indicators, combine the patient's historical health records and lifestyle information, analyze the warning signals of human health, and generate the dynamic health evaluation results; Using the dynamic health evaluation results, analyze the association between the recorded health indicator data and the potential health risks, identify the risk factors, evaluate the development trend and impact of the potential health risks, and generate the health risk analysis results.

10. A human health early warning system based on physical examination data analysis, characterized in that, According to the method for analyzing human health warnings based on physical examination data according to any one of claims 1-9, the system includes: The time sorting module sorts the physical examination data points based on the historical physical examination data of human blood pressure, blood sugar, cholesterol, and body mass index, and generates the time series sorting results; The spatial layout evaluation module analyzes the distribution of the physical examination data points in a finite-dimensional space based on the time series sorting results, identifies the areas with health abnormalities, and marks them to generate the warning level mapping results; The risk clustering module conducts clustering analysis on the health indicators of the physical examination data points based on the warning level mapping results, identifies and marks the groups with the same health risk characteristics, and generates the health risk group identification results; The risk factor association module analyzes the physical examination data in the differentiated groups based on the health risk group identification results, captures the correlation between the health indicators, and identifies the key health risk factors by comparing the differences and relationships of the health indicators to generate the associated health risk factors; The health trend prediction module analyzes the health warning signals in the future time period based on the associated health risk factors, identifies the potential health trends, and analyzes the changes and impacts of the health risks to generate the health risk analysis results.

Images