Health management method and system based on data analysis

Abstract

The invention relates to the technical field of data processing, in particular to a health management method and system based on data analysis, and the method comprises the steps: obtaining health data, setting a label, constructing a historical database, adding the health data of a to-be-tested person into data with a normal label under the corresponding age of the historical database, an anomaly detection data set is constructed and preprocessed; obtaining the relative density of each group of data in the preprocessed anomaly detection data set, obtaining the anomaly confidence of each parameter in each group of data of all abnormal health data in all health data, obtaining a local outlier correction coefficient, weighting the local outlier of the LOF algorithm according to the local outlier correction coefficient, and obtaining the abnormal health data of all health data; and evaluating the body health condition of the health data of the to-be-tested person according to the weighted local outlier degree. According to the method, a more accurate evaluation result is obtained according to the weighted local outlier degree in the LOF algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing. More specifically, the present invention relates to a health management method and system based on data analysis. Background Art

[0002] Health management is crucial in modern society. It mainly evaluates the health status and prevents the occurrence of diseases by analyzing an individual's physical data. Indicators such as height, weight, and body fat percentage are key factors in health management. Abnormal weight or body fat percentage may indicate health problems such as obesity and malnutrition, which in turn affect cardiovascular health, metabolic function, etc. Therefore, accurately evaluating the distribution of these indicators in the normal population and thereby determining whether an individual has health risks is of great significance for formulating personalized health management plans.

[0003] The Chinese patent application document with the publication number CN119181462A discloses a health management method with health data analysis function. The health management method with health data analysis function includes the following steps: S1. Collect health data information of the user's body through existing devices and establish a personal health record based on these data information. The body data information includes but is not limited to blood pressure, heart rate, and blood sugar; S2. The collected health data needs to be cleaned, sorted, and preprocessed before establishing a personal health record for subsequent data analysis; S3. According to the user's health data, use machine learning models and deep learning model algorithms to conduct health risk assessment and predict possible health risks of the user, including but not limited to cardiovascular diseases and diabetes.

[0004] The health management method with health data analysis function provided in this application document can improve the accuracy and personalization level of health management and meet people's health needs. However, since the collected health data approximately follows different distributions at different ages, there are differences in the data concentration regions, and the abnormal ranges of health data at different ages are different. When directly using the LOF algorithm to evaluate health, the outlier degree scales in each region are different and the abnormal credibility is different, which will lead to inaccurate calculation results of local outlier degree, and further make the subsequent health assessment results inaccurate. Summary of the Invention

[0005] To solve the problem that when directly using the LOF algorithm to evaluate health for different ages approximately following different distributions, the outlier degree scales in each region are different and the abnormal credibility is different, resulting in inaccurate calculation results of local outlier degree, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a health management method based on data analysis includes: obtaining health data, sorting the body data, setting tags, constructing a historical database, adding the health data of the person to be tested to the data with the tag of normal under the corresponding age in the historical database, constructing an anomaly detection dataset and performing preprocessing; obtaining the relative density of each group of data in the preprocessed anomaly detection dataset, where the relative density is used to reflect the possibility of the abnormal state of each group of data; obtaining the anomaly confidence of each parameter in each group of data with abnormal health data among all the health data of the person with the age of The anomaly confidence is used to identify the outliers in the data; calculating the local outlier factor correction coefficient of the health data of the person to be tested, weighting the local outlier factor of the LOF algorithm according to the local outlier factor correction coefficient, and evaluating the physical health condition of the health data of the person to be tested according to the weighted local outlier factor.

[0007] The effect is that: by comprehensively analyzing the relative density, anomaly confidence, local outlier factor and local outlier factor correction coefficient of health data, a precise, dynamic and personalized method for evaluating the physical health of the person to be tested is provided, which not only improves the recognition accuracy of abnormal situations, but also helps doctors and researchers make more accurate diagnosis and intervention decisions, thus promoting the healthy growth of the person to be tested.

[0008] Preferably, the constructing the anomaly detection dataset and performing preprocessing includes: The anomaly detection dataset includes: historical data with the tag of normal and the data of the person to be tested with the age of Taking any data group in the health data of the person with the age of in the anomaly detection dataset as the target group, where the target group is sorted in the order of height, weight, and body fat percentage in the health data; Performing maximum-minimum normalization processing on the target group to obtain the numerical size of each parameter after normalization.

[0009] The effect is that: by constructing an anomaly detection dataset including historical data with the tag of normal and the data of the person to be tested with the age of and performing ordered sorting and maximum-minimum normalization processing on the health data (height, weight, and body fat percentage in the target group, and in addition to the above conventional data, other data such as waist circumference, blood pressure, skinfold thickness, palm length, etc. can also be collected according to needs) in the target group, the beneficial effect of this method is to provide a standardized and comparable data environment, thereby enhancing the accuracy and efficiency of anomaly detection. The normalization processing ensures the consistency between different parameters, makes the evaluation results more objective and reliable, and helps to quickly identify possible abnormal health data.

[0010] Preferably, obtaining the relative density of each group of data in the anomaly detection data set comprises the steps of: According to any data group in the anomaly detection data set as the target group, calculate the age of the anomaly detection data set under the square root: The sum of squares of the differences between each parameter of the target group and each parameter of other groups except the target group is calculated to obtain the Euclidean distance between the target group and other groups, and the Euclidean distance is used as the spacing; Among them, the anomaly detection dataset represents age , and have a total of Group data, calculate all The average value of the distance between the target group and other groups except the target group in the group data, and the database age is calculated as The label is normal The sum of the distances between the target group and other groups except the target group in the group data, and the ratio of the average value to the sum of the distances is used as the age in the anomaly detection data set The relative density of the target group data.

[0011] The effect is: by calculating the relative density between the target data group and other groups, abnormal data groups can be accurately identified, thereby improving the data anomaly detection capability; by analyzing the difference in parameters of different data groups, key features that have a greater impact on anomaly detection can be identified, thereby improving the quality and efficiency of data analysis.

[0012] Preferably, the age is obtained as The abnormal confidence of each parameter in each set of data of all health data abnormalities, including: Among them, the age in the anomaly detection dataset is The total number of health data is , separate the healthy data from the anomaly detection dataset. The total number of normal healthy data is The total number of abnormal health data is , calculate the mean of the parameter data in all abnormal health data, and use the mean as the abnormal confidence of each parameter.

[0013] The effect is: by calculating the mean of various parameters in the abnormal health data as the abnormality confidence, it helps to accurately identify abnormal indicators and assess the physical health risks of the tested persons.

[0014] Preferably, the local outlier correction coefficient includes: Separate healthy data from anomaly detection dataset The total number of normal healthy data is , calculate the relative density of the health data of the tested person in the anomaly detection data set and the relative density of the age of the tested person in the anomaly detection data set The absolute difference between the relative densities of the respective data regions, and normalize the absolute difference by dividing it by the average of the sum of the distances from the health data of the person to be tested to the respective data regions, to obtain a relative density deviation value; Use the ratio between the abnormal confidence level of the health data of the person to be tested and the relative density deviation value as a local outlier correction coefficient.

[0015] Preferably, the weighting of the local outlier factor of the LOF algorithm according to the local outlier correction coefficient includes: Use the product of the local outlier correction coefficient and the local outlier factor of the LOF algorithm as the weighted local outlier factor.

[0016] Preferably, the assessment of the physical health condition of the health data of the person to be tested includes: Obtain the weighted local outlier factor of the health data of the person to be tested in the abnormal detection dataset, set an outlier threshold, and in response to the weighted local outlier factor being greater than the outlier threshold, it indicates that there is an abnormality in the physical health of the person to be tested.

[0017] In a second aspect, a health management system based on data analysis includes: a processor and a memory, and the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned health management method based on data analysis is implemented.

[0018] The present invention has the following effects: 1. The present invention respectively obtains the relative densities and abnormal confidence levels of the health data regions of different ages according to the database, and then obtains the local weight of the LOF algorithm to weight the reachable density, and evaluates the physical health condition of the health data of the person to be tested according to the weighted local outlier factor, so as to obtain a more accurate assessment result of the physical health of the person to be tested.

[0019] 2. By comprehensively considering multiple parameters in the health data and calculating their abnormal confidence levels and relative densities, the present invention can more comprehensively evaluate the physical health status of the person to be tested. Comprehensively evaluate the abnormalities of the person to be tested in multiple health data indicators, thereby improving the accuracy and reliability of the evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 is a flowchart of the method of steps S1 - S3 in the health management method based on data analysis according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0022] The following will describe in detail the specific implementation manners of the present invention in conjunction with the accompanying drawings.

[0023] Referring to Figure 1 , the health management method based on data analysis includes steps S1 - S3, specifically as follows: S1: Obtain health data, sort the body data, set labels, construct a historical database, add the health data of the person to be tested to the data with the label of normal under the corresponding age in the historical database, and construct an anomaly detection data set and perform preprocessing.

[0024] Further explanation, use a high-precision height measuring instrument to measure the height of the person to be tested. Require the person to be tested to stand at attention, feet together, shoulders relaxed, head upright, eyes looking straight ahead, and measure the vertical distance from the vertex to the ground; use a weighing scale to measure the weight of the person to be tested, ensuring that the measurement is carried out without shoes and heavy clothes; use a professional body fat scale to measure the body fat percentage of the person to be tested. When measuring, the person to be tested stands steadily on the scale surface, hands hanging naturally, and the body remains balanced. The data collected is not limited to the above, and can be collected according to the actual situation.

[0025] In addition to the above conventional data, other data can also be collected according to needs, such as waist circumference, blood pressure, skin fold thickness, palm length, etc. After the measurement is completed, record and store the measurement results in the computer, sort the various parameters of the health data in the order of height, weight, and body fat percentage, and record the collected health data as: , the age of the person to be tested is .

[0026] Specifically, further explain the historical database. By collecting a large amount of health data of different ages historically and obtaining the physical health status according to the physical examination results of the person to be tested, set labels, use the letter to represent the physical health status, represents abnormal, represents normal. Construct a historical database according to the health data of different ages and the corresponding physical health status, and filter out the abnormal data with missing items to ensure the integrity and effectiveness of the data used in the subsequent processing. Denote all the health data of age in the health database after filtering out the missing values as the The value of the th parameter of the group of data is , the physical health condition is , and the age is . The total number of health data is , among which the total number of health data with normal health data is , and the total number of health data parameters is .

[0027] Since the health data of the person to be tested is added to the data with the label of normal under the corresponding age in the historical database to construct an anomaly detection data set , the anomaly detection data set includes: historical data with the label of normal and data of the person to be tested with the age of . Taking any data group in the health data with the age of in the anomaly detection data set as the target group, where the target group is sorted in the order of height, weight, and body fat percentage in the health data; Perform maximum-minimum normalization processing on the target group to obtain the numerical size of each parameter after normalization .

[0028] S2: Obtain the relative density of each group of data in the preprocessed anomaly detection data set. The relative density is used to reflect the possibility of the abnormal state of each group of data; obtain the anomaly confidence of each parameter in each group of data with abnormal health data among all health data with the age of . The anomaly confidence is used to identify the outliers in the data.

[0029] Taking any data group in the anomaly detection data set as the target group, calculate the sum of the squares of the differences between each parameter of the target group with the age of in the anomaly detection data set and each parameter of other groups except the target group to obtain the Euclidean distance between the target group and other groups, and use the Euclidean distance as the spacing; Among them, the anomaly detection data set represents the age of , and there are a total of groups of data. Calculate the average value of the spacing between the target group and other groups except the target group in all groups of data, calculate the sum of the spacing between the target group and other groups except the target group in the groups of data with the label of normal in the database with the age of , and use the ratio of the average value and the sum of the spacing as the relative density of the target group data with the age of in the anomaly detection data set.

[0030] Specifically, the relative density satisfies the following relational expression: ; In the formula, Indicates the relative density of the th group of data with age in the anomaly detection dataset, Indicates the distance between the th group of data with age in the anomaly detection dataset and the th group of data except the Indicates the total number of healthy data with normal healthy data among the healthy data with age .

[0031] Furthermore, it should be noted that when the distance between the target group of data in the anomaly detection dataset and other groups of data except the target group of data is higher, it indicates that the density of this group of data area is smaller.

[0032] According to the relative density, it should be noted that the possibility of different age healthy data being labeled as abnormal in different regions is different. For regions where different age healthy data have a high frequency of abnormal labels, it indicates that when the corresponding age healthy data is in this region, it is likely to be labeled as abnormal, so the anomaly confidence level in this region is high; when the corresponding age group of healthy data is in a region with a low frequency of abnormal labels, it indicates that the tested personnel with body data in this region are not likely to be labeled as abnormal, so the anomaly confidence level in this region is low.

[0033] Among them, the total number of healthy data with age in the anomaly detection dataset is , the total number of healthy data with normal healthy data separated from the anomaly detection dataset is and the total number of healthy data with abnormal healthy data is . Calculate the mean value of the parameter data among all healthy data with abnormal healthy data, and use the mean value as the anomaly confidence level of each parameter.

[0034] In addition, another embodiment further includes: Specifically, obtain the mean value of the anomaly confidence levels of three parameter data in the healthy data as the anomaly confidence level, where the three parameters include but are not limited to: height, weight, and body fat percentage.

[0035] Specifically, the degree of anomaly satisfies the following relational expression: ; In the formula, represents the anomaly confidence level of the healthy data of the tested personnel, represents the total number of healthy data with age , represents the total number of healthy data with normal healthy data among the healthy data with age , represents the age of The total number of health data with abnormal health data for all Indicates the value of the th parameter in the health data of the person to be tested Indicates that the age is Among all the health data with abnormal health data, the th group of data The value of the th parameter

[0036] It should be noted that Indicates the abnormal confidence level of the th parameter of the collected health data; Indicates the average abnormal confidence level of all parameters of the collected health data; When the parameters of the health data of the person to be tested are closer to the data with abnormal health data as a whole, the abnormal confidence level of the health data of the person to be tested is high, and the value tends to 1; when the parameters of the health data of the person to be tested are farther away from the data with abnormal health data as a whole, the abnormal confidence level of the health data of the person to be tested is low, and the value tends to 0.

[0037] S3: Calculate the local outlier factor correction coefficient of the health data of the person to be tested , weight the local outlier factor of the LOF algorithm according to the local outlier factor correction coefficient, and evaluate the physical health status of the health data of the person to be tested according to the weighted local outlier factor.

[0038] The local outlier factor correction coefficient includes: The total number of health data with normal health data separated from the abnormal detection dataset is , calculate the absolute difference between the relative density of the health data of the person to be tested in the abnormal detection dataset and the relative density of each group of data regions with age in the abnormal detection dataset, and normalize the absolute difference by dividing it by the average value of the sum of the distances from the health data of the person to be tested to each group of data to obtain the relative density deviation value; Take the ratio between the abnormal confidence level of the health data of the person to be tested and the relative density deviation value as the local outlier factor correction coefficient.

[0039] Specifically, the local outlier factor correction coefficient satisfies the following relational expression: ; In the formula, Represents the local outlier factor correction coefficient of the health data of the person to be tested, Represents the abnormal confidence level of the health data of the person to be tested, Represents that the age is The total number of health data with normal health data in the health data represents the relative density of the health data of the person to be tested in the anomaly detection dataset, represents the relative density of the th group of data regions with age in the anomaly detection dataset, represents the distance from the health data of the person to be tested to the th group of data, represents the normalization function.

[0040] It should be noted that represents the average rate of change of the relative density of the collected health data region with respect to distance; when the average rate of change of the relative density of the collected health data region with respect to distance is high and the anomaly confidence level is low, the possibility of an anomaly is low, and the local outlier factor correction coefficient of the collected health data is low; when the average rate of change of the relative density of the collected health data region with respect to distance is low and the anomaly confidence level is high, the possibility of an anomaly is high, and the local outlier factor correction coefficient of the collected health data is high.

[0041] Multiply the local outlier factor correction coefficient by the local outlier factor of the LOF algorithm to obtain the weighted local outlier factor.

[0042] Specifically, the weighted local outlier factor satisfies the following relational expression: ; In the formula, represents the weighted local outlier factor of the health data of the person to be tested in the anomaly detection dataset, represents the local outlier factor correction coefficient of the health data of the person to be tested, represents the local outlier factor of the health data of the person to be tested obtained by the LOF algorithm in the anomaly detection dataset.

[0043] Obtain the weighted local outlier factor of the health data of the person to be tested in the anomaly detection dataset, set the outlier threshold, and when the weighted local outlier factor is greater than the outlier threshold, it indicates that there is an anomaly in the physical health of the person to be tested.

[0044] For further explanation, the outlier threshold , when the weighted local outlier factor , it is considered that there is an anomaly in the physical health of the person to be tested, and a comparative analysis is performed with the health data of those with good physical health to obtain the body fat percentage and weight management plan.

[0045] The specific steps are as follows: Obtain all the health data of people with good physical health of the same age as the person to be tested and record it as the control dataset, obtain the Euclidean distance between the health data of the person to be tested and each data in the control dataset, and set the empirical nearest neighbor number , select the data with the closest Euclidean distance to the health data of the person to be tested in the control dataset as the management standard data. Take the mean of each parameter of all management standard data as the body fat percentage and weight standard of the person to be tested, and output the health problems of the person to be tested and the body fat percentage and weight standard of the person to be tested to the report form, and timely remind the medical staff to assist the person to be tested in managing the body fat percentage and weight according to the corresponding standards.

[0046] An embodiment of the present invention also discloses a health management system based on data analysis. The system includes a processor and a memory. The memory stores computer program instructions. When the computer program instructions are executed by the processor, the health management method based on data analysis according to the first aspect of the present invention is implemented.

[0047] The system also includes a communication bus, a communication interface and other components well known to those skilled in the art. Their settings and functions are known in the art, so they will not be described in detail here.

[0048] In the present invention, the aforementioned memory can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application program, module, or both. Any such computer storage medium can be part of the device or accessible or connectable to the device.

[0049] In the description of this specification, "a plurality of" and "several" mean at least two, such as two, three or more, unless otherwise specifically defined.

[0050] Although this specification has shown and described multiple embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many variations, modifications, and alternative means will occur to those skilled in the art without departing from the spirit and scope of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in practicing the present invention.

Claims

1. A health management method based on data analysis, characterized in that: include: Obtain health data, sort the health data, set labels, build a historical database, add the health data of the person to be tested to the data labeled as normal at the corresponding age in the historical database, build anomaly detection data set and perform preprocessing; Get the relative density of each group of data in the anomaly detection data set after preprocessing, and the relative density is used to reflect the possibility of abnormal state of each group of data; get the age abnormal confidence of each parameter in each group of data of all abnormal health data in the health data, wherein the abnormal confidence is used to identify abnormal values ​​in the data; Calculate the local outlier correction coefficient for the health data of the tested person , the local outlier degree of the LOF algorithm is weighted according to the local outlier degree correction coefficient, and the physical health condition of the health data of the person to be tested is evaluated according to the weighted local outlier degree.

2. The health management method based on data analysis according to claim 1, characterized in that: The construction of anomaly detection data set and preprocessing include: The anomaly detection dataset includes: The labels are normal historical data and the data of the person to be tested, with the age in the anomaly detection dataset as Any data group in the health data is taken as the target group, wherein the target group is sorted in the order of height, weight, and body fat percentage in the health data; The target group is normalized to obtain the normalized value of each parameter.

3. The health management method based on data analysis according to claim 1, characterized in that: Obtaining the relative density of each group of data in the anomaly detection data set includes the steps of: According to any data group in the anomaly detection data set as the target group, calculate the age of the anomaly detection data set under the square root: The square sum of the differences between each parameter of the target group and each parameter of other groups except the target group is obtained to obtain the Euclidean distance between the target group and other groups, and the Euclidean distance is used as the spacing; wherein the anomaly detection dataset represents age , and have a total of Group data, calculate all The average value of the distance between the target group and other groups except the target group in the group data, and the database age is calculated as The label is normal The sum of the distances between the target group and other groups except the target group in the group data, and the ratio of the average value to the sum of the distances is used as the age in the anomaly detection data set The relative density of the target group data.

4. The health management method based on data analysis according to claim 1, characterized in that: Get the age as The abnormal confidence of each parameter in each group of data of all health data abnormalities, include: Among them, the age in the anomaly detection dataset is The total number of health data is , separate the healthy data from the anomaly detection dataset. The total number of normal healthy data is The total number of abnormal health data is , calculate the mean of the parameter data in all abnormal health data, and use the mean as the abnormal confidence of each parameter.

5. The health management method based on data analysis according to claim 1, characterized in that: The local outlier correction coefficient includes: Separate healthy data from anomaly detection dataset The total number of normal healthy data is , calculate the relative density of the health data of the tested person in the anomaly detection data set and the relative density of the age of the tested person in the anomaly detection data set The absolute difference between the relative densities of each group of data areas is normalized by dividing the absolute difference by the average value of the sum of the distances from the health data of the tested person to each group of data to obtain the relative density deviation value; The ratio between the abnormal confidence level and the relative density deviation value of the health data of the tested person is taken as the local outlier correction coefficient.

6. The health management method based on data analysis according to claim 1, characterized in that: The step of weighting the local outlier degree of the LOF algorithm according to the local outlier degree correction coefficient includes: The product of the local outlier correction coefficient and the local outlier of the LOF algorithm is taken as the weighted local outlier.

7. The health management method based on data analysis according to claim 1, characterized in that: Evaluate the physical health of the person's health data, including: The weighted local outlier degree of the health data of the person to be tested in the anomaly detection data set is obtained, and an outlier threshold is set. When the weighted local outlier degree is greater than the outlier threshold, the person to be tested has an abnormality in physical health.

8. A health management system based on data analysis, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a health management method based on data analysis according to any one of claims 1 to 7 is implemented.

Images