Resident health risk evaluation method and system based on multi-factor index and entropy weight method

Abstract

The invention belongs to the technical field of public safety, and provides a resident health risk evaluation method and system based on multi-factor indexes and an entropy weight method.The method comprises the steps that principal component analysis and dimension reduction processing are conducted on a plurality of factor indexes related to the environment and a plurality of factor indexes related to residents; performing multi-factor variance analysis by taking whether health risks exist or not as dependent variables and taking the principal components subjected to dimension reduction processing as independent variables to obtain interaction among factors; and determining the weight of each principal component index after dimension reduction processing by using an entropy weight method, and performing weighted summation on each principal component index to obtain a health comprehensive evaluation index. And finally, carrying out risk evaluation on the health of the residents by utilizing the interaction among the factors and the comprehensive health evaluation indexes. The method comprehensively considers a plurality of factor indexes related to the external environment and a plurality of factor indexes related to residents, can reflect the real situation of the risk of influencing the health of the residents, and guarantees the evaluation precision while reducing the data processing amount.

Description

Technical Field

[0001] The present invention belongs to the field of public safety technology, and in particular relates to a resident health risk assessment method and system based on multi-factor indicators and entropy weight method. Background Art

[0002] With the acceleration of urban construction and industrialization, global warming, and increasingly severe air pollution, high temperature environments and air pollutants are seriously affecting residents' health.

[0003] At present, the concentration of air pollutants and high temperature exposure in a certain area are usually used to evaluate the health risks of residents. This only takes into account the impact of the external environment on residents' health, but does not consider the impact of residents' own conditions on residents' health. There is a lack of comprehensive evaluation indicators related to the external environment and residents' own factors, which leads to low accuracy in the health risk assessment of residents in the region and cannot reflect the real situation that affects residents' health risks. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a resident health risk assessment method and system based on multi-factor indicators and entropy weight method. The present invention first performs principal component analysis and dimensionality reduction processing on multiple factor indicators related to the environment and multiple factor indicators related to the residents themselves; then takes the presence or absence of health risk as the dependent variable and the principal component after dimensionality reduction processing as the independent variable, and performs multi-factor variance analysis to obtain the interaction between factors; and, uses the entropy weight method to determine the weight of each principal component indicator after dimensionality reduction processing, and performs weighted summation of each principal component indicator to obtain a comprehensive health evaluation index; finally, uses the interaction between factors and the comprehensive health evaluation index to evaluate the risk affecting the health of residents; comprehensively considers multiple factor indicators related to the external environment and multiple factor indicators related to the residents themselves, can reflect the real situation affecting the health risk of residents, and ensures the evaluation accuracy while reducing the amount of data processing.

[0005] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0006] In a first aspect, the present invention provides a method for evaluating residents' health risks based on multi-factor indicators and entropy weight method, comprising:

[0007] Obtain multiple factor indicators related to the environment in the evaluation area, as well as multiple factor indicators related to the residents themselves in the evaluation area;

[0008] Conduct principal component analysis on all factor indicators to obtain the principal component data that affect residents' health status; perform dimensionality reduction on the principal component data;

[0009] The presence or absence of health risk was used as the dependent variable, and the principal component after dimensionality reduction was used as the independent variable to conduct multi-factor variance analysis to obtain the interaction between factors; and the entropy weight method was used to determine the weight of each principal component index after dimensionality reduction, and the weighted sum of each principal component index was obtained to obtain the comprehensive health evaluation index; the weight was optimized and adjusted by comparing and verifying the eigenvalues ​​obtained by principal component analysis with the weight values, and the comprehensive health evaluation index was corrected by using the interaction results of the interaction;

[0010] The interaction between factors and the revised comprehensive health evaluation indicators are used to evaluate the risks that affect residents' health.

[0011] Furthermore, the multiple factor indicators related to the environment include temperature index, air pollutant concentration, normalized difference vegetation index, gross domestic product and population; the multiple factor indicators related to the residents themselves include age, gender, body mass index, education level, residential area and environment, income, eating habits and health status.

[0012] Furthermore, when performing principal component analysis, the factors are taken as column vectors and arranged into a matrix of preset rows and preset columns by columns, and the average value of the corresponding row is subtracted from each variable in the matrix to obtain a new matrix; the eigenvalues ​​and eigenvectors of the covariance matrix of the new matrix are calculated, the eigenvalues ​​are sorted from large to small, and the eigenvalues ​​are matched one by one with the eigenvectors, and the principal components with a preset number of eigenvalues ​​are selected according to the size of the eigenvalues;

[0013] When the weight value of the factor does not correspond to the size of the eigenvalue of the principal component, it means that the data is inaccurate; when the weight value of the factor and the eigenvalue of the principal component differ by only a preset number of data, the eigenvalue is processed and the average value is taken with the weight, where the processed eigenvalue is the ratio of the eigenvalue to the sum of the eigenvalues; when the weight value of the factor corresponds to the eigenvalue of the principal component, the entropy weight method is directly used to obtain the weight for calculation.

[0014] Furthermore, when conducting interaction analysis, the interaction effects between factors are first determined through multiplicative interaction, and then decomposed through additive interaction; the impact on health status Z is:

[0015] Z=gA+hB+jC+lD+mA′(aB+bC)

[0016] Among them, A′(aB+bC) represents multiplicative interaction; A, B, C and D are population, pollutant concentration, temperature and age, respectively; a and b are multiplicative interaction coefficients; g, h, j, l and m are additive interaction coefficients.

[0017] Furthermore, the relative excess risk caused by the interaction in the additive interaction is equal to the risk of both the first factor and the second factor minus the risk of the first factor not existing and the second factor existing, minus the risk of the first factor existing and the second factor not existing, plus 1; the relative excess risk caused by the interaction in the multiplicative interaction is equal to the ratio of the risk of both the first factor and the second factor existing to the risk of the first factor not existing and the second factor existing, divided by the risk of the first factor existing and the second factor not existing; when the relative excess risk caused by the interaction in the additive interaction is greater than 0, the risk of the factors acting together is greater than the sum of their individual effects; when the relative excess risk caused by the interaction in the additive interaction is less than 0, the risk of the factors acting together is less than the sum of their individual effects; when the relative excess risk caused by the interaction in the additive interaction is equal to 0, the risk of the factors acting together is additive; when the relative excess risk caused by the interaction in the multiplicative interaction is greater than 1, the risk of the factors acting together is greater than the product of their individual effects, otherwise, the risk of the factors acting together is less than the product of their individual effects;

[0018] The principal components after dimensionality reduction are standardized; the standardized principal components are converted into a standardized data matrix; the weight of each indicator in the standardized data matrix is ​​calculated to construct a weight matrix; the entropy weight of each indicator is calculated based on the weight matrix, and the difference coefficient is calculated based on the entropy weight; the weight of each indicator is calculated based on the difference coefficient, and the comprehensive health evaluation index of residents is obtained through weight coupling.

[0019] Furthermore, the relationship between two factors and the influence of multiple factors on another factor are obtained through interaction; if the two factors have a positive interaction, a correction coefficient greater than 1 is introduced; if the two factors have a negative interaction, a correction coefficient less than 1 is introduced; otherwise, no correction coefficient is introduced; the entropy weight method is used to obtain the weights corresponding to temperature, pollutants, body mass index and eating habits. The principal component analysis method shows that there is no obvious interaction between the four factors and the three factors, and no new correction coefficient is introduced; the result of the two-factor interaction is that there is a positive interaction between temperature and pollutants, and between body mass index and eating habits. Two correction coefficients greater than 1 are introduced, and the revised comprehensive health evaluation index RHI is:

[0020] RHI=m(αT+βP)+n(γBMI+λH)

[0021] Among them, T is temperature; P is pollutant; BMI is body mass index; H is eating habits; α, β, γ and λ are weights; m and n are two correction coefficients greater than 1.

[0022] In a second aspect, the present invention also provides a resident health risk assessment system based on multi-factor indicators and entropy weight method, comprising:

[0023] The data collection module is configured to: obtain multiple factor indicators related to the environment in the evaluation area, and multiple factor indicators related to the residents themselves in the evaluation area;

[0024] The data dimension reduction module is configured to: perform principal component analysis on all factor indicators to obtain principal component data that affect residents' health status; perform dimension reduction processing on the principal component data;

[0025] The comprehensive evaluation index determination module is configured as follows: taking the presence or absence of health risk as the dependent variable and the principal component after dimensionality reduction as the independent variable, performing multi-factor variance analysis to obtain the interaction between factors; and using the entropy weight method to determine the weight of each principal component index after dimensionality reduction, performing weighted summation of each principal component index to obtain the comprehensive health evaluation index; using the comparison and verification of the eigenvalues ​​obtained by the principal component analysis and the weight values, optimizing and adjusting the weights, and using the interaction results of the interaction to correct the comprehensive health evaluation index;

[0026] The evaluation module is configured to use the interaction between factors and the modified comprehensive health evaluation indicators to conduct risk assessment on residents' health.

[0027] In a third aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for evaluating resident health risks based on multi-factor indicators and entropy weight method described in the first aspect.

[0028] In a fourth aspect, the present invention also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein when the processor executes the program, the steps of the method for evaluating resident health risks based on multi-factor indicators and entropy weight method described in the first aspect are implemented.

[0029] In a fifth aspect, the present invention also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the method for evaluating resident health risks based on multi-factor indicators and entropy weight method described in the first aspect.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] In the present invention, principal component analysis and dimensionality reduction are first performed on multiple factor indicators related to the environment and multiple factor indicators related to the residents themselves; then, the presence or absence of health risks is taken as the dependent variable, and the principal components after dimensionality reduction are taken as the independent variables, and multi-factor variance analysis is performed to obtain the interaction between factors; and, the entropy weight method is used to determine the weights of each principal component indicator after dimensionality reduction, and the weighted sum of each principal component indicator is obtained to obtain a comprehensive health evaluation index; finally, the interaction between factors and the comprehensive health evaluation index are used to evaluate the risks affecting the health of residents; the multiple factor indicators related to the external environment and the multiple factor indicators related to the residents themselves are comprehensively considered, which can reflect the real situation affecting the health risks of residents. At the same time, through principal component analysis, dimensionality reduction, optimization and adjustment of weights, and correction of the comprehensive health evaluation index using the interaction results, the evaluation accuracy is guaranteed on the basis of reducing the amount of data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0033] Figure 1 is a flow chart of the method of embodiment 1 of the present invention;

[0034] Figure 2 This is the principal component analysis and dimensionality reduction process of Example 1 of the present invention;

[0035] Figure 3 The multi-factor variance analysis process of Example 1 of the present invention;

[0036] Figure 4 This is the entropy weight method analysis process of Example 1 of the present invention;

[0037] Figure 5 The evaluation coupling process of Example 1 of the present invention;

[0038] Figure 6 The weight optimization of embodiment 1 of the present invention;

[0039] Figure 7 The correction coefficient determination process of Example 1 of the present invention;

[0040] Figure 8 This is a system structure block diagram of embodiment 2 of the present invention. DETAILED DESCRIPTION

[0041] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0042] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0043] Embodiment 1:

[0044] With the acceleration of industrialization, global warming, and increasing air pollution, high temperature and air pollutants have seriously affected residents' health. In addition, age, living habits, place of residence, body mass index (BMI), education level, income level, etc. also have potential impacts on residents' health. Therefore, it has become more urgent to identify and select key factors from many health risk factors and to control them in a targeted manner.

[0045] In the existing technology, air pollutant concentrations and high temperature exposure are usually used to evaluate residents' health risks. However, there are still many factors such as age, living habits, place of residence, BMI, education level, income level, etc. that have an impact on residents' health, and the types of evaluation factors are insufficient. In addition, when evaluating the proportion of residents' health risks of various factors, the existing methods lack evaluation of a specific factor, and cannot achieve targeted control of residents' health risks in a certain area or age group. There is also a lack of relevant comprehensive indicators to evaluate residents' health risks.

[0046] Therefore, existing technologies lack multi-factor considerations in health risk assessment, and the control of key factors that affect residents' health under specific factors is not perfect, which reduces the reliability of residents' health risk assessment.

[0047] Based on this, this embodiment provides a resident health risk assessment method based on multi-factor indicators and entropy weight method. First, relevant data is collected, then dimensionality reduction is performed through principal component analysis, and then the interaction effect between factors is obtained through multi-factor variance analysis and interpreted and evaluated. Finally, the weight of each indicator is calculated through entropy weight method and coupled to the comprehensive evaluation index of resident health. The specific steps are:

[0048] S1. Obtain data such as temperature index, air pollutant concentration, Normalized Difference Vegetation Index (NDVI), gross domestic product (GDP), and population of multiple pre-selected cities through reasonable and legal means. Also, obtain data such as age, gender, body mass index, education level, residential area and environment, income, eating habits, and health status of preset or random populations in multiple pre-selected urban areas through random questionnaires and other methods.

[0049] The optional way to obtain data is to obtain basic data of multiple cities or regions, as well as the population in the region, through online inquiries, on-site measurements, and questionnaire surveys, etc., to provide data for subsequent multi-factor interactive evaluation. Among them, temperature indicators can be obtained through relevant meteorological information centers if permitted; pollutants, urban GDP and population can be obtained through relevant statistical websites if permitted; the normalized vegetation index can be obtained through the geospatial data cloud GEOSS; and relevant population data can be obtained through questionnaire surveys.

[0050] Optionally, the temperature index is a city in a certain year, which is higher than 95% of the maximum temperature for at least 2 days, 3 days or 4 days in the warm season. The warm season can be determined according to different regions. For example, the warm season in a certain city is from June 1 to September 30.

[0051] Optionally, the categorical variables are indexed, where the temperature variable is indexed as 1, 2 or 3 for days above 95% of the maximum temperature for at least 2, 3 or 4 days; PM2.5 is indexed as less than 75 mg / m 3 Excellent, 75 to 150 mg / m 3 Good, greater than 150 mg / m 3 The corresponding variable indexes are 1, 2 and 3 for pollution; NDVI is divided into sparse green plants between 0 and 0.2 and dense green plants above 0.2, and the corresponding variable indexes are 0 and 1; the noise environment is divided into a quiet environment of less than 60dB, a noisy environment of 60 to 100dB, and a severe noise environment of more than 100dB, and the corresponding variable indexes are 1 to 3; according to age, it is divided into children under 12 years old, young people between 12 and 35 years old, middle-aged people between 36 and 60 years old, and the elderly over 60 years old, and the corresponding indicator variables are 1 to 4; according to urban population, cities are divided into four classification variables: small cities with a population of less than 2 million, medium-sized cities with a population of 2 million to 5 million, large cities with a population of 5 million to 10 million, and megacities with a population of more than 10 million. , and its corresponding indicator variable is 1 to 4; the indicator variables corresponding to gender are 0 and 1 for male and female; the education level is divided into primary school education or below, junior high school education, high school education, and college education or above, and the corresponding indicator variables are 1 to 4; BMI is divided into underweight (less than 18.5), normal weight (18.5 to 24.9), overweight (25 to 29.9), and obesity (greater than 30), and the corresponding indicator variables are 1 to 4; the indicator variable for high-salt and spicy diet is 1, and the indicator variable for light diet is 0; the variable indicator for having the habit of staying up late is 1, and the variable indicator for not staying up late is 0; the variable indicator for high work pressure is 1, and the variable indicator for low work pressure is 0; the indicator variable for smoking or drinking is 1, and the indicator variable for not smoking or drinking is 0; the indicator variable for having health risks is 1, and the indicator variable for not having health risks is 0.

[0052] Optionally, the Normalized Difference Vegetation Index (NDVI) can be obtained through the geospatial data cloud GEOSS. The NDVI value ranges from -1 to 1. The closer to 1, the denser the vegetation cover, the closer to 0, the bare land, and the closer to -1, the snow or lake. The calculation formula is:

[0053]

[0054] Among them, NIR represents the radiation value of the near-infrared band; Red represents the radiation value of the red light band.

[0055] S2. Using principal component analysis (PCA), the principal components that have a greater impact on residents' health status are obtained from the data in step S1 to perform data dimensionality reduction.

[0056] Optionally, in principal component analysis, the factors are treated as column vectors and arranged into a matrix of n rows and m columns. The mean value of each row is subtracted from each variable in the matrix to obtain a new matrix X. This approach is called decentralization, which puts all data at the same reference point and avoids the impact of differences in the mean values ​​of each feature on subsequent calculations.

[0057] Calculate the eigenvalues ​​and eigenvectors of the covariance matrix of matrix X, sort the eigenvalues, and correspond them one by one with the corresponding eigenvectors, and select k principal components according to the size of the eigenvalues. Among them, the matrix X is:

[0058]

[0059] Among them, X i is the average value of a row, and the calculation formula is:

[0060]

[0061] The formula for calculating the covariance matrix is:

[0062]

[0063] Calculate the eigenvalues ​​l1, l2…l according to the covariance matrix m And sort them from large to small, select the vectors corresponding to the first k eigenvalues ​​as the principal components to achieve the dimensionality reduction effect, where k is less than m.

[0064] In one of the embodiments, through the principal component analysis method, it can be found that factors such as continuous high temperature, air pollution, age, eating habits, smoking and drinking, work pressure, and staying up late have higher eigenvalues, while the eigenvalues ​​of population size, education level and income status for health risks are smaller. For education level and income status, it may be because some of the people corresponding to the questionnaire are young people and children who are receiving education and have no income or have very low income and are in good health. The impact of the interaction between the two on health risks can be determined in the multivariate analysis of variance.

[0065] S3. Perform a multi-factor analysis of variance (Factorial ANOVA) on the data after dimensionality reduction in step S2, and evaluate and interpret the results.

[0066] Optionally, in multifactor ANOVA, SPSS (Statistical Package for the Social Sciences) can be used to take the presence or absence of health risk as the dependent variable and other principal components as independent variables, select a univariate model for multifactor ANOVA, obtain the interaction and P value between single factors or multiple factors, and finally interpret and evaluate the results.

[0067] Among them, select the single variable in the linear model, select the presence or absence of health risk as the dependent variable, select the other factors as fixed factors, select the full factor model, and the significance level can be selected as 0.05.

[0068] The P value indicates the probability that the result is caused by chance. It is used to observe whether the data is extreme and takes a value between 0 and 1. The interaction includes additive interaction and multiplicative interaction. The principle and calculation formula of additive interaction are as follows:

[0069] RERI=RR 11 -RR 01 -RR 10 +1

[0070] Among them, RERI is the relative excess risk caused by the interaction; RR is the risk ratio, RR 11 The risk that both the first and second factors exist; RR 01 The risk that the first factor does not exist but the second factor exists; RR 10 The risk that the first factor exists and the second factor does not exist; 1 represents RR 00 ;

[0071] For multivariate RERI, taking three elements as an example, the calculation formula is:

[0072] RERI=RR 111 -RR 110 -RR 101 -RR011 +RR 100 +RR 010 +RR 001 -2RR 000

[0073] Among them, the subscript 1 represents existence, and 0 represents non-existence. When RERI is greater than 0, it is a positive interaction, indicating that the risk of the combined effect of the factors is greater than the sum of their individual effects; when RERI is less than 0, it is a negative interaction, indicating that the risk of the combined effect of the factors is less than the sum of their individual effects; when RERI is equal to 0, there is no interaction, indicating that the risk of the combined effect of the factors is additive.

[0074] The principle of multiplicative interaction and its calculation formula are:

[0075]

[0076] When Y is greater than 1, it is a positive interaction, indicating that the risk of the combined effect of the factors is greater than the product of their individual effects; when Y is less than 1, it is a negative interaction, indicating that the risk of the combined effect of the factors is less than the product of their individual effects. The hazard ratio is:

[0077]

[0078] Here, i, j = 1, 0, where 1 indicates that the situation occurs, and 0 indicates that the situation does not occur.

[0079] Additive interaction and multiplicative interaction are two different methods. Additive interaction assumes that the impact of each factor on the whole can be directly added, while multiplicative interaction assumes that each factor is independent of each other. Nevertheless, additive interaction can be converted to multiplicative interaction through logarithms. For example, Z=A+B+C can be converted through ln(Z)=ln(A)+ln(B)+ln(C)=ln(ABC). Additive interaction is more suitable for situations where the impact can be linearly superimposed, focusing on the impact of individual factors. Multiplicative interaction is more suitable for situations where there is a synergistic effect between factors, emphasizing the interaction between factors. The interaction effect between factors can be determined first through multiplicative interaction, and then a more detailed decomposition can be obtained through additive interaction.

[0080] For example, consider the impact of urban population, pollutant concentration, temperature and age on health status, and define these factors as A, B, C, D and Z respectively. Through multiplicative interaction, it is found that urban population has strong multiplicative interaction with pollutant concentration and temperature respectively, that is, changes in population will affect pollutant concentration or temperature. If there is no multiplicative interaction among the three, then A′(aB+bC) can be used to represent the multiplicative interaction, where a and b are the corresponding multiplicative interaction coefficients. The larger the interaction coefficient, the more significant the multiplicative interaction. Next, calculate the additive interaction, that is, Z=gA+hB+jC+lD+mA′(aB+bC), where g, h, j, l and m are additive interaction coefficients respectively. The larger the coefficient, the more significant the additive interaction. When multiplicative interaction and additive interaction act together, it is also possible to consider the impact of only a few factors on health status. For example, considering the impact of age and temperature on health status, it can be expressed as Z=jC+lD. If considering the impact of population, age and temperature on health status, it can be expressed as Z=gA+jC+lD+mA′bC.

[0081] Any observation value is the result of deviation from the mean. The deviation is mainly caused by random errors and treatment effects. Therefore, any observation value is the sum of the mean, random errors and treatment effects. The calculation formula is:

[0082] x ij =μ+ε+αβ

[0083] Among them, μ is the mean; ε is the random error and the within-group effect; αβ is the treatment effect and the between-group effect; if the treatment effect is significantly larger than the random error, it means that the effect of the interaction treatment is significant, and a significant P value will be obtained, that is, P is close to 0; if the treatment effect is similar to or even smaller than the random error, it means that the interaction effect has little impact on the result, and an insignificant P value will be obtained, that is, P is larger, which means that the probability of the corresponding data occurring is smaller.

[0084] Optionally, the evaluation of the data is based on a comprehensive consideration of the output results and research needs. When considering the impact of a single factor on health risks, the single factor interaction can be viewed, and the impact of the factor on health risks can be judged based on the interaction results; when considering the impact of multiple factors on health risks, first obtain the main effects of the independent variables through SPSS software, and then check the interaction between the independent variables. If there is an interaction, it is necessary to explain why the two are interdependent, and evaluate the impact on the dependent variable based on positive or negative interaction; if there is a large interaction effect between the independent variable and the dependent variable, then it is necessary to consider taking certain targeted measures for single or multiple dependent variables to reduce residents' health risks.

[0085] In one of the embodiments, multifactor analysis of variance shows that there is a large single-factor interaction between continuous high temperature, pollutants, work pressure, staying up late, smoking and drinking, high-salt diet, etc. and health risks, which is basically similar to the principal components with higher eigenvalues ​​obtained in the principal component analysis method; and there is also a high interaction between work pressure and income, and staying up late, and there is a large P value. At this time, this situation is not statistically significant, and it is impossible to judge the impact of the three factors on health risks. Other multifactor evaluations are similar, and the specific situation is analyzed according to actual cases.

[0086] S4. Calculate the weight of each evaluation index through the entropy weight method, and calculate the comprehensive evaluation index of residents' health through weight coupling.

[0087] Optionally, in the entropy weight, the original data of the principal component is standardized to form a standardized matrix, the weight of each indicator is calculated, a weight matrix is ​​constructed, the entropy weight and the coefficient of difference of each indicator are calculated, the weight of each indicator is calculated, and the comprehensive evaluation index of residents' health is obtained through weight coupling.

[0088] S4.1. Standardize the data of the principal component, eliminate the dimension effect, and unify the data range. The calculation formula is:

[0089]

[0090] Among them, x j is the j-th index value; x min is the minimum value among the j-th index values; x max is the maximum value of the j-th index value; r ij is the standardized index value.

[0091] For example, if there are k evaluation indicators and n sets of data, the standardized principal component is converted into a standardized data matrix R = (r ij ) k′n , of the following form:

[0092]

[0093] S4.2. Calculate the weight m of the i-th item of the j-th indicator ij , and its calculation formula is:

[0094]

[0095] S4.2. Construct a weight matrix Q by weighting each indicator. ij} k′n , which has the following form:

[0096]

[0097] S4.2. Calculate the entropy weight e of the jth indicator based on the jth column element of the weight matrix Q j , the entropy weight calculation formula is:

[0098]

[0099] Among them, the constant This is to ensure that the entropy value is between 0 and 1.

[0100] S4.3. Calculate the coefficient of variation d to represent the degree of non-uniformity of the indicator. The calculation formula is:

[0101] d j =1-e j

[0102] The coefficient of difference directly affects the size of the weight. The larger the coefficient of difference, the greater its importance to the evaluation and the greater its weight.

[0103] S4.4. Weight w of the jth indicator j The calculation formula is:

[0104]

[0105] Among them, the weight w j The larger the value, the greater the effect of the corresponding j-th independent variable on the dependent variable.

[0106] S4.5. Combine the j-th indicator and the j-th weight w j Multiply and sum to obtain the comprehensive evaluation index of residents' health, which is used to evaluate the impact on residents' health. The calculation formula is as follows:

[0107]

[0108] Among them, RHI is a comprehensive evaluation index of residents’ health. The larger the RHI, the greater the health risk of the residents.

[0109] Among them, the larger the weight value of a factor calculated by the entropy weight method, the more obvious the impact of the factor on health. In the principal component analysis, the eigenvalue corresponding to the factor should also be large. Therefore, the weight calculated by the entropy method can be compared with the corresponding principal component eigenvalue in the principal component analysis to verify the accuracy of the experimental model. By using the entropy weight method, it can be concluded that continuous high temperature, BMI, age and pollutant concentration have high weight values, and noise, eating habits, staying up late, smoking and drinking, etc. also have high weight values. Specifically, by comparing and verifying the weight value and the eigenvalue, the weight value can be optimized and adjusted; if the weight value of the factor cannot correspond to the size of the principal component eigenvalue, it means that the data may be inaccurate; if the weight value of the factor and the principal component eigenvalue are different only in individual data, the processed eigenvalue and weight can be averaged; if the weight value of the factor and the principal component eigenvalue can correspond well, the weight obtained by the entropy weight method is directly taken for calculation; among them, the calculation formula of the processed eigenvalue x′ is as follows:

[0110]

[0111] Optionally, the entropy weight method can be performed through EXCEL. First, enter the principal component index in the EXCEL list; secondly, obtain the weight matrix according to the standardization formula and the weight formula, and list them in the workbook; then obtain the difference coefficient d of the corresponding index through the entropy weight calculation formula and the difference coefficient calculation formula j , and are listed below the corresponding weight matrix; finally, the weight value w of the corresponding indicator is calculated through the weight calculation formula j , and listed in the corresponding coefficient of difference d j Below, it is used to couple the evaluation indicators into corresponding resident health evaluation indicators.

[0112] Interaction is used to understand the relationship between variables and explain the level at which the effect of one factor depends on the effect of another factor. Through interactive evaluation, the relationship between two factors and the impact of multiple factors on another factor can be obtained. In this embodiment, coupling factors to obtain the comprehensive health evaluation index is a summation process, so only the impact of additive interaction is considered. If two factors have a positive interaction, it means that the combined effect of the two factors has a great impact on health, so a correction coefficient greater than 1 is introduced. When the interaction between the two factors is not obvious, no new correction coefficient needs to be introduced. If two factors have a negative interaction, it means that the combined effect of the two factors has a small impact on health, so a correction coefficient less than 1 is introduced.

[0113] The entropy weight method was used to obtain the weights of temperature T, pollutant P, BMI, and dietary habits H, which are α, β, γ, and λ. The principal component analysis method showed that there was no obvious interaction between the four factors and the three factors, so no new correction coefficients were needed in these cases. The two-factor interaction results showed that temperature T and pollutant P, BMI and dietary habits H all had positive interactions. Two correction coefficients m and n greater than 1 were introduced. If temperature T and pollutant P had a stronger positive interaction, m would be greater than n. The final calculation formula for the comprehensive evaluation index of residents' health RHI was:

[0114] RHI=m(αT+βP)+n(γBMI+λH)

[0115] Among them, the optional numerical value selection method of the correction coefficients m and n is: the interaction results are graded, when there is a negative interaction, the result is less than -1, and the correction coefficient is 0; when the interaction result is between -1 and -0.9, the correction coefficient is 0.1, and by analogy, when it is between -0.2 and -0.1, the correction coefficient is 0.9; similarly, when there is a positive interaction, the result is greater than 1, and the correction coefficient is 2; the interaction result is 0.9 to 1, and the correction coefficient is 1.9; and by analogy, when it is 0.1 to 0.2, the correction coefficient is 1.1; therefore, when the interaction result is between -0.1 and 0.1, it is determined that the interaction effect is small and no correction coefficient is introduced.

[0116] In summary, this embodiment first performs dimensionality reduction processing on data such as heat waves, PM2.5, NVDI, and age through the PCA principal component analysis method, and obtains the factors with the largest eigenvalues ​​as the principal components; secondly, the principal components are used as independent variables and the presence or absence of health risks are used as dependent variables to obtain the interaction between single factors or multiple factors and the P values ​​of different situations through the multifactor variance analysis method to determine the impact on health risks and whether they are statistically significant, and to perform a reasonable interpretation and evaluation; then the k principal components are standardized, and the weight values ​​of each indicator are obtained through the entropy weight method, and the weight values ​​are compared and verified with the eigenvalues ​​of the corresponding principal components to verify the accuracy of the model, and the comprehensive evaluation index of residents' health is coupled through the weight value; finally, the interpretation evaluation obtained by the multifactor variance analysis method and the comprehensive health evaluation index obtained by the entropy weight method are used to perform an index-based, diversified, and systematic comprehensive evaluation of health risks, and targeted countermeasures to reduce health risks are proposed.

[0117] Embodiment 2:

[0118] This embodiment provides a resident health risk assessment system based on multi-factor indicators and entropy weight method, including:

[0119] The data collection module is configured to: obtain multiple factor indicators related to the environment in the evaluation area, and multiple factor indicators related to the residents themselves in the evaluation area;

[0120] The data dimension reduction module is configured to: perform principal component analysis on all factor indicators to obtain principal component data that affect residents' health status; perform dimension reduction processing on the principal component data;

[0121] The comprehensive evaluation index determination module is configured as follows: taking the presence or absence of health risk as the dependent variable and the principal component after dimensionality reduction as the independent variable, performing multi-factor variance analysis to obtain the interaction between factors; and using the entropy weight method to determine the weight of each principal component index after dimensionality reduction, performing weighted summation of each principal component index to obtain the comprehensive health evaluation index; using the comparison and verification of the eigenvalues ​​obtained by the principal component analysis and the weight values, optimizing and adjusting the weights, and using the interaction results of the interaction to correct the comprehensive health evaluation index;

[0122] The evaluation module is configured to use the interaction between factors and the modified comprehensive health evaluation indicators to conduct risk assessment on residents' health.

[0123] The working method of the system is the same as the resident health risk assessment method based on multi-factor indicators and entropy weight method in Example 1, and will not be repeated here.

[0124] Embodiment 3:

[0125] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method for evaluating the health risks of residents based on multi-factor indicators and entropy weight method described in Example 1 are implemented.

[0126] Embodiment 4:

[0127] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the program, the steps of the method for evaluating resident health risks based on multi-factor indicators and entropy weight method described in Example 1 are implemented.

[0128] Embodiment 5:

[0129] This embodiment provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the method for evaluating residents' health risks based on multi-factor indicators and entropy weight method described in Example 1 are implemented.

[0130] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. The residents' health risk assessment method based on multi-factor indicators and entropy weight method is characterized by: include: Obtain multiple factor indicators related to the environment in the evaluation area, as well as multiple factor indicators related to the residents themselves in the evaluation area; Conduct principal component analysis on all factor indicators to obtain the principal component data that affect residents' health status; perform dimensionality reduction on the principal component data; The presence or absence of health risk was used as the dependent variable, and the principal component after dimensionality reduction was used as the independent variable to conduct multi-factor variance analysis to obtain the interaction between factors; and the entropy weight method was used to determine the weight of each principal component index after dimensionality reduction, and the weighted sum of each principal component index was obtained to obtain the comprehensive health evaluation index; the weight was optimized and adjusted by comparing and verifying the eigenvalues ​​obtained by principal component analysis with the weight values, and the comprehensive health evaluation index was corrected by using the interaction results of the interaction; The interaction between factors and the revised comprehensive health evaluation indicators are used to evaluate the risks that affect residents' health.

2. The resident health risk assessment method based on multi-factor indicators and entropy weight method as claimed in claim 1 is characterized in that: The multiple factor indicators related to the environment include temperature index, air pollutant concentration, normalized difference vegetation index, gross domestic product and population; the multiple factor indicators related to the residents themselves include age, gender, body mass index, education level, residential area and environment, income, eating habits and health status.

3. The resident health risk assessment method based on multi-factor indicators and entropy weight method as claimed in claim 1 is characterized in that: When performing principal component analysis, the factors are taken as column vectors and arranged into a matrix with preset rows and columns. The mean value of the corresponding row is subtracted from each variable in the matrix to obtain a new matrix. Calculate the eigenvalues ​​and eigenvectors of the covariance matrix of the new matrix, sort the eigenvalues ​​from large to small, and make a one-to-one correspondence with the eigenvectors, select the principal components that are in the front of a preset number according to the size of the eigenvalues, and obtain the principal component eigenvalues; When the weight value of the factor does not correspond to the size of the eigenvalue of the principal component, it means that the data is inaccurate; if the weight value of the factor and the eigenvalue of the principal component differ by only a preset number of data, the eigenvalue is processed and the average value is taken with the weight, where the processed eigenvalue is the ratio of the eigenvalue to the sum of the eigenvalues; When the weight value of the factor corresponds to the eigenvalue of the principal component, the entropy weight method is directly used to obtain the weight for calculation.

4. The resident health risk assessment method based on multi-factor indicators and entropy weight method as claimed in claim 1 is characterized in that: When conducting interaction analysis, the interaction effects between factors are first determined through multiplicative interaction, and then decomposed through additive interaction; the impact on health status Z is: Z=gA+hB+jC+lD+mA′(aB+bC) Among them, A′(aB+bC) represents multiplicative interaction; A, B, C and D are population, pollutant concentration, temperature and age, respectively; a and b are multiplicative interaction coefficients; g, h, j, l and m are additive interaction coefficients.

5. The resident health risk assessment method based on multi-factor indicators and entropy weight method as claimed in claim 4 is characterized in that: In an additive interaction, the relative excess risk caused by the interaction is equal to the risk of both the first and second factors minus the risk of the first factor not existing but the second factor existing, minus the risk of the first factor existing but the second factor not existing, plus 1; in a multiplicative interaction, the relative excess risk caused by the interaction is equal to the ratio of the risk of both the first and second factors existing to the risk of the first factor not existing but the second factor existing, divided by the risk of the first factor existing but the second factor not existing; when the relative excess risk caused by the interaction in an additive interaction is greater than 0, the risk of the factors acting together is greater than the sum of their individual effects; When the relative excess risk caused by the interaction in the additive interaction is less than 0, the risk when the factors act together is less than the sum of their individual effects; When the relative excess risk caused by the interaction in the additive interaction is equal to 0, the risk of the joint action between the factors is additive; when the relative excess risk caused by the interaction in the multiplicative interaction is greater than 1, the risk of the joint action between the factors is greater than the product of their individual actions, otherwise, the risk of the joint action between the factors is less than the product of their individual actions; Standardize the principal components after dimensionality reduction; Convert the standardized principal components into a standardized data matrix; Calculate the weight of each indicator in the standardized data matrix and construct a weight matrix; calculate the entropy weight of each indicator based on the weight matrix, and calculate the difference coefficient based on the entropy weight; The weight of each indicator is calculated according to the coefficient of difference, and the comprehensive health evaluation index of residents is obtained through weight coupling.

6. The resident health risk assessment method based on multi-factor indicators and entropy weight method as claimed in claim 1 is characterized in that: The relationship between two factors and the influence of multiple factors on another factor are obtained through interaction. If the two factors have a positive interaction, a correction coefficient greater than 1 is introduced; if the two factors have a negative interaction, a correction coefficient less than 1 is introduced. Otherwise, no correction factor is introduced; The entropy weight method was used to obtain the weights corresponding to temperature, pollutants, body mass index and eating habits. The principal component analysis method showed that there was no obvious interaction between the four factors and the three factors, and no new correction coefficient was introduced. The two-factor interaction result showed that there was a positive interaction between temperature and pollutants, and between body mass index and eating habits. Two correction coefficients greater than 1 were introduced, and the revised comprehensive health evaluation index RHI was: RHI=m(αT+βP)+n(γBMI+λH) Among them, T is temperature; P is pollutant; BMI is body mass index; H is eating habits; α, β, γ and λ are weights; m and n are two correction coefficients greater than 1.

7. The resident health risk assessment system based on multi-factor indicators and entropy weight method is characterized by: include: The data collection module is configured to: obtain multiple factor indicators related to the environment in the evaluation area, and multiple factor indicators related to the residents themselves in the evaluation area; The data dimension reduction module is configured to: perform principal component analysis on all factor indicators to obtain principal component data that affect residents' health status; perform dimension reduction processing on the principal component data; The comprehensive evaluation index determination module is configured as follows: taking the presence or absence of health risk as the dependent variable and the principal component after dimensionality reduction as the independent variable, performing multi-factor variance analysis to obtain the interaction between factors; and using the entropy weight method to determine the weight of each principal component index after dimensionality reduction, performing weighted summation of each principal component index to obtain the comprehensive health evaluation index; using the comparison and verification of the eigenvalues ​​obtained by the principal component analysis and the weight values, optimizing and adjusting the weights, and using the interaction results of the interaction to correct the comprehensive health evaluation index; The evaluation module is configured to use the interaction between factors and the modified comprehensive health evaluation indicators to conduct risk assessment on residents' health.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the resident health risk assessment method based on multi-factor indicators and entropy weight method as described in any one of claims 1 to 6 are implemented.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the program, the steps of the resident health risk assessment method based on multi-factor indicators and entropy weight method as described in any one of claims 1-6 are implemented.

10. A computer program product, characterized in that The computer program product includes a computer program, which, when executed by a processor, implements the steps of the method for evaluating resident health risks based on multi-factor indicators and entropy weight method as described in any one of claims 1 to 6.

Images