Method for determining weight of evaluation index of health-care type tourism resort

Abstract

The invention provides a weight determination method for evaluation indexes of a health-care type tourism resort. The method comprises the following steps: constructing a target-criterion-index three-layer system; collecting a multi-expert judgment matrix by an AHP method, and introducing confidence coefficient weighted average to obtain a comprehensive matrix; calculating and verifying a preliminary weight; the dynamic threshold consistency check iteration is passed; and performing multi-stage synthesis, defuzzification and iterative correction by using a fuzzy comprehensive evaluation method to obtain a final weight. According to the method, the scientificity and accuracy of weight determination of the evaluation indexes of the health-care tourism resort can be improved, and a reliable basis is provided for evaluation and optimization of the resort.

Description

Technical Field

[0001] This invention relates to the field of evaluation technology for health and wellness tourism resorts, and more specifically, to a method for determining the weights of evaluation indicators for health and wellness tourism resorts. Background Technology

[0002] With the improvement of people's living standards, health and wellness tourism resorts are receiving increasing attention. Currently, the evaluation of these resorts mainly relies on simple indicator systems and subjective judgments, lacking scientific and systematic methods for determining weights. Existing evaluation methods mostly employ single methods such as the Analytic Hierarchy Process (AHP) or the Fuzzy Comprehensive Evaluation (FCE), but these methods have limitations. For example, AHP is rather crude in determining expert weights, failing to fully consider multi-dimensional factors such as experts' industry experience and professional background, and its consistency test threshold is fixed, making dynamic adjustments based on indicator importance and data distribution impossible. FCE, in revising weights, lacks an effective iterative mechanism, making it difficult to ensure the accuracy and stability of the final weights. Furthermore, existing methods often neglect data availability and the importance of expert review when constructing evaluation indicator systems, leading to the selection of some indicators being insufficiently scientific and reasonable.

[0003] In implementing the embodiments of the present invention, the prior art has at least the following problems or defects: the existing evaluation methods cannot comprehensively consider multi-dimensional factors to determine expert weights, the threshold for consistency testing is fixed and cannot be dynamically adjusted, and the construction of the indicator system lacks scientificity and rationality, resulting in insufficient accuracy and reliability of the evaluation results. Summary of the Invention

[0004] This invention provides a method for determining the weights of evaluation indicators for health and wellness tourism resorts, including: Step S1: Establish a hierarchical structure for the evaluation index system of health and wellness tourism resorts. This hierarchical structure includes an objective layer, a criterion layer, and an indicator layer. The criterion layer is divided into multiple criteria based on predefined rules, and the indicator layer determines specific indicators based on data availability and expert review. Step S2: Based on the Analytic Hierarchy Process (AHP), construct pairwise comparison judgment matrices for multiple experts on the evaluation indicators, and introduce expert confidence coefficients. Calculate expert weights based on each expert's industry experience value, professional background value, years of service value, professional title level value, and historical evaluation accuracy value. Perform a weighted average of multiple judgment matrices based on the expert weights to obtain a comprehensive judgment matrix. Step S3: Based on the comprehensive judgment matrix, calculate the preliminary weight vector of the evaluation index and verify the effectiveness of the preliminary weight vector; Step S4: Perform a consistency check on the comprehensive judgment matrix, calculate the consistency ratio, and dynamically adjust the qualified threshold of the consistency ratio according to the importance of the indicators and the data distribution. Iterate until the consistency ratio is less than or equal to the qualified threshold. Step S5: Use the fuzzy comprehensive evaluation method (FCE) to correct the initial weight vector, including constructing a fuzzy evaluation matrix, determining the evaluation set, performing multi-level fuzzy synthesis operations, and adjusting the initial weight vector according to the synthesis results to obtain the final evaluation index weights.

[0005] Furthermore, establishing the hierarchical structure in step S1 specifically includes: Step S1.1: Define the target layer as a comprehensive evaluation of health and wellness tourism resorts; Step S1.2: Based on literature analysis and expert interviews, the criteria layer is divided into four criteria: environmental quality, service facilities, health and wellness activities and management services. Among them, the environmental quality criteria include sub-criteria for natural environment and human environment, and the service facilities criteria include sub-criteria for infrastructure and special health and wellness facilities. Step S1.3: Set an indicator layer for each sub-criteria. The indicators for the natural environment sub-criteria include air quality index, water quality index and vegetation coverage. The indicators for the human environment sub-criteria include richness of cultural resources and community friendliness. The indicators for the infrastructure sub-criteria include convenient transportation and comfortable accommodation. The indicators for the health and wellness facilities sub-criteria include coverage of medical care points and proportion of healthy dining restaurants. Step S1.4: Through expert voting and data analysis, remove indicators with a data missing rate higher than the preset threshold and retain the final indicator layer.

[0006] Furthermore, the calculation of expert weights in step S2 specifically includes: Step S2.1: Using the 1-9 scaling method, each expert independently constructs a judgment matrix; Step S2.2: Calculate the expert confidence coefficient for each expert. This coefficient is based on industry experience value, professional background value, years of practice value, professional title level value, and historical evaluation accuracy value. The historical evaluation accuracy value is obtained by comparing the consistency between the expert's historical evaluation and the actual results. Step S2.3: Normalize the expert confidence coefficients to obtain the expert weight for each expert; Step S2.4: If any expert's expert weight is lower than the preset weight threshold, then exclude that expert and recalculate the expert weight; Step S2.5: Use expert weights to perform a weighted arithmetic average on multiple judgment matrices to obtain a comprehensive judgment matrix.

[0007] Furthermore, the formula for calculating the expert confidence coefficient in step S2.2 is as follows:

[0008] in, This represents the expert confidence coefficient of the j-th expert. This represents the normalized industry experience value. This represents the normalized professional background value. This represents the normalized value of years of service. This represents the normalized professional title grade value. This represents the normalized historical evaluation accuracy value. The weights are predefined weights, and satisfy the following conditions: Among them, the historical evaluation accuracy value The correlation coefficient was calculated by comparing historical evaluations by experts with real data.

[0009] Furthermore, step S3, which involves calculating the initial weight vector and verifying its effectiveness, specifically includes: Step S3.1: Calculate the largest eigenvalue and its corresponding eigenvector of the comprehensive judgment matrix using the eigenvector method; Step S3.2: Normalize the feature vectors to obtain the preliminary weight vectors; Step S3.3: Check if there are negative or zero values ​​in the initial weight vector. If so, return to step S2 to reconstruct the judgment matrix. Step S3.4: Calculate the variance of the initial weight vector. If the variance is lower than the preset variance threshold, then smooth the weights.

[0010] Furthermore, the qualified threshold for dynamically adjusting the consistency ratio in step S4 specifically includes: Step S4.1: Calculate the consistency index CI and consistency ratio CR of the comprehensive judgment matrix, where CI is calculated using the formula... Calculate, where, CR is the largest eigenvalue of the comprehensive judgment matrix, where n is the order of the comprehensive judgment matrix; CR is expressed by the formula... Calculate, where RI is the average random consistency index; Step S4.2: Set the initial pass threshold and adjust the threshold according to the importance of the indicator. The pass threshold for important indicators is lowered by 0.02 to 0.05, and the pass threshold for non-important indicators is raised by 0.01 to 0.03. Step S4.3: Analyze the data distribution of the index values, calculate the variance and skewness of each index, and if the variance is higher than the preset variance threshold and the skewness exceeds the preset skewness range, then further adjust the qualified threshold. Step S4.4: If the adjusted CR is still greater than the qualified threshold, return to step S2 to recalculate the expert weights or reconstruct the judgment matrix.

[0011] Furthermore, the classification of indicator importance in step S4.2 is based on expert ratings and the indicator's position in the hierarchical structure, specifically including: Step S4.2.1: Experts score the indicators on a scale of 1 to 5, and calculate the average score for each indicator; Step S4.2.2: Calculate the importance score of the indicator by combining the hierarchical depth of the indicator in the criterion layer and the sub-criterion layer; Step S4.2.3: Based on the importance score, the indicators are divided into three categories: high importance, medium importance, and low importance, and different qualified threshold adjustment ranges are set for each category.

[0012] Furthermore, the correction using the fuzzy comprehensive evaluation method (FCE) in step S5 specifically includes: Step S5.1: Determine the evaluation set as five levels, including excellent, good, average, poor, and very poor; Step S5.2: Collect expert evaluations of each indicator in the indicator layer and construct a fuzzy evaluation matrix R, where the elements... This represents the proportion of experts who gave the j-th level evaluation for the i-th indicator; Step S5.3: Perform multi-level fuzzy synthesis operation. First, perform primary synthesis on the indicators within the indicator layer, and then perform secondary synthesis on the criteria layer to obtain the fuzzy evaluation result vector B. Step S5.4: Defuzzify the fuzzy evaluation result vector B, and use the weighted average method to obtain the comprehensive score vector S for each indicator; Step S5.5: Iteratively correct the initial weight vector W using the comprehensive score vector S until the weight change is less than the preset convergence threshold.

[0013] Furthermore, the multi-level fuzzy synthesis operation in step S5.3 specifically includes: Step S5.3.1: For each criterion, perform initial synthesis using the preliminary weight vector W and the fuzzy evaluation matrix R, calculated as follows: ,in, This is the fuzzy evaluation result of the k-th criterion. It is the initial weight sub-vector corresponding to the k-th criterion. It is the fuzzy evaluation submatrix corresponding to the k-th criterion, with the symbol... Represents a fuzzy composition operator; Step S5.3.2: Convert the initial synthesis result As input for the second-level synthesis, the weights of the criterion layer are combined to perform the second-level synthesis, resulting in the overall fuzzy evaluation result vector B.

[0014] Furthermore, the iterative correction in step S5.5 specifically includes: Step S5.5.1: Set the initial weights to the initial weight vector W, and the iteration counter t=0; Step S5.5.2: Calculate the comprehensive score vector S under the current weight; Step S5.5.3: Use the formula Update the weights, where, It is the weight of the i-th indicator in the t-th iteration. It is the weight of the i-th indicator in the (t+1)-th iteration. It is the comprehensive score of the i-th indicator. It is the weight of the k-th indicator in the t-th iteration. It is the comprehensive score of the k-th indicator, and n is the total number of indicators; Step S5.5.4: Calculate the weight change ,in, It is the change in weight. It is the weight of the i-th indicator in the (t+1)-th iteration. It is the weight of the i-th indicator in the t-th iteration, where n is the total number of indicators; if If the number of iterations is less than the preset convergence threshold or the maximum number of iterations is reached, the iteration stops and the final weight is output; otherwise, t=t+1 and return to step S5.5.2.

[0015] The embodiments of the present invention have at least the following beneficial effects: 1. By constructing a hierarchical structure comprising a target layer, a criterion layer, and an indicator layer, and combining the Analytic Hierarchy Process (AHP) with the Fuzzy Comprehensive Evaluation (FCE) method, this invention can comprehensively consider multi-dimensional factors and scientifically determine the weights of evaluation indicators for health and wellness tourism resorts. This method not only considers experts' industry experience, professional background, years of experience, professional title, and historical evaluation accuracy, but also ensures the scientific validity and accuracy of the evaluation indicator weights by dynamically adjusting the acceptable threshold for consistency ratios. This effectively solves the problems of existing technologies having a single method for determining weights and lacking scientific basis, thus improving the credibility of the evaluation results.

[0016] 2. By introducing and normalizing expert confidence coefficients, this invention can calculate more reasonable expert weights based on the multi-dimensional characteristics of experts, avoiding evaluation biases caused by individual expert differences. Simultaneously, through iterative loops and dynamic adjustment of the consistency ratio threshold, this invention ensures the consistency of the comprehensive judgment matrix, further improving the accuracy and reliability of the evaluation index weights. This effectively solves the problems of coarse expert weight determination and fixed consistency test thresholds in existing technologies, enhancing the scientific rigor and adaptability of the evaluation method.

[0017] 3. This invention employs a fuzzy comprehensive evaluation method to correct the initial weight vector. By combining multi-level fuzzy evaluations of indicators by experts, and through multi-level synthesis operations and iterative corrections, a final weight that better reflects reality is obtained. This method not only considers the importance of indicators and data distribution but also ensures the stability and accuracy of the evaluation indicator weights by dynamically adjusting the acceptable threshold for the consistency ratio. This effectively solves the problems of unreasonable indicator system construction and lack of effective correction mechanisms in existing technologies, improving the scientific rigor and practicality of the evaluation results. Attached Figure Description

[0018] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation, wherein: Figure 1 This is a flowchart illustrating a method for determining the weights of evaluation indicators for health and wellness tourism resorts, as provided in an embodiment of the present invention. Detailed Implementation

[0019] The principles and spirit of the invention will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are provided merely to enable those skilled in the art to better understand and implement the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided to make the invention more thorough and complete, and to fully convey the scope of the invention to those skilled in the art.

[0020] Those skilled in the art will recognize that embodiments of the present invention can be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0021] The number of any elements in the accompanying drawings is for illustrative purposes only and not as a limitation, and any naming is for distinction only and has no limiting meaning.

[0022] The following is for reference. Figure 1 , Figure 1 This is a flowchart illustrating a method for determining the weights of evaluation indicators for health and wellness tourism resorts, as provided in an embodiment of the present invention. Figure 1 As shown, a method for determining the weights of evaluation indicators for health and wellness tourism resorts includes: Step S1: Establish a hierarchical structure for the evaluation index system of health and wellness tourism resorts. This hierarchical structure includes an objective layer, a criterion layer, and an indicator layer. The criterion layer is divided into multiple criteria based on predefined rules, and the indicator layer determines specific indicators based on data availability and expert review. Step S2: Based on the Analytic Hierarchy Process (AHP), construct pairwise comparison judgment matrices for multiple experts on the evaluation indicators, and introduce expert confidence coefficients. Calculate expert weights based on each expert's industry experience value, professional background value, years of service value, professional title level value, and historical evaluation accuracy value. Perform a weighted average of multiple judgment matrices based on the expert weights to obtain a comprehensive judgment matrix. Step S3: Based on the comprehensive judgment matrix, calculate the preliminary weight vector of the evaluation index and verify the effectiveness of the preliminary weight vector; Step S4: Perform a consistency check on the comprehensive judgment matrix, calculate the consistency ratio, and dynamically adjust the qualified threshold of the consistency ratio according to the importance of the indicators and the data distribution. Iterate until the consistency ratio is less than or equal to the qualified threshold. Step S5: Use the fuzzy comprehensive evaluation method (FCE) to correct the initial weight vector, including constructing a fuzzy evaluation matrix, determining the evaluation set, performing multi-level fuzzy synthesis operations, and adjusting the initial weight vector according to the synthesis results to obtain the final evaluation index weights.

[0023] This invention discloses a method for determining the weights of evaluation indicators for health and wellness tourism resorts. This method scientifically determines the weights of evaluation indicators by establishing a hierarchical structure comprising a target layer, a criterion layer, and an indicator layer. The target layer represents a comprehensive evaluation of the health and wellness tourism resort; the criterion layer comprises multiple criteria based on literature analysis and expert interviews; and the indicator layer consists of specific indicators determined based on data availability and expert review. The Analytic Hierarchy Process (AHP) is a method for multi-criterion decision-making, which determines the relative importance of each indicator by constructing a judgment matrix. The Fuzzy Comprehensive Evaluation (FCE) method is an evaluation method based on fuzzy mathematics, capable of handling fuzziness and uncertainty in the evaluation process, and correcting weights through fuzzy evaluation and comprehensive calculation of the indicators.

[0024] In step S1, the target layer is defined as a comprehensive evaluation of health and wellness tourism resorts, which is the ultimate goal of the evaluation system. The criteria layer is divided into four criteria: environmental quality, service facilities, health and wellness activities, and management services. Each criterion is further subdivided into sub-criterions; for example, the environmental quality criterion includes sub-criterions for natural environment and human environment. The indicator layer consists of specific evaluation indicators, such as air quality index, water quality index, and vegetation coverage rate under the natural environment sub-criterion. The selection of these indicators is based on data availability and expert review, ensuring their scientific validity and practicality.

[0025] In step S2, the calculation of expert weights considers multiple dimensions, including industry experience, professional background, years of experience, professional title, and historical evaluation accuracy. The weight for each expert is obtained through normalization. The judgment matrix is ​​constructed using a 1-9 proportional scaling method, completed independently by each expert, reflecting their judgment on the relative importance of each indicator.

[0026] In step S3, the initial weight vector is calculated using the eigenvector method. This is achieved by calculating the largest eigenvalue of the comprehensive judgment matrix and its corresponding eigenvector, followed by normalization. Validity verification includes checking for negative or zero values ​​in the weight vector and calculating the variance to assess the weight distribution.

[0027] Furthermore, in step S2, the expert confidence coefficient is calculated by comprehensively considering multiple dimensions, including the expert's industry experience, professional background, years of experience, professional title, and historical evaluation accuracy. These parameters are obtained from the expert's personal data and historical evaluation data, and are normalized to ensure they are comparable under the same scale. Weighting coefficients are predefined and used to balance the contributions of different factors.

[0028] In step S4, the consistency ratio is calculated by comparing the relationship between the largest eigenvalue of the comprehensive judgment matrix and the matrix order. When dynamically adjusting the pass threshold, the threshold is adjusted according to the importance of the indicators; the pass threshold for important indicators is appropriately lowered, while the pass threshold for less important indicators is appropriately raised. Furthermore, the data distribution of the indicator values ​​is analyzed, and the variance and skewness of each indicator are calculated to further adjust the pass threshold.

[0029] In some embodiments, establishing the hierarchical structure in step S1 specifically includes: Step S1.1: Define the target layer as a comprehensive evaluation of health and wellness tourism resorts; Step S1.2: Based on literature analysis and expert interviews, the criteria layer is divided into four criteria: environmental quality, service facilities, health and wellness activities and management services. Among them, the environmental quality criteria include sub-criteria for natural environment and human environment, and the service facilities criteria include sub-criteria for infrastructure and special health and wellness facilities. Step S1.3: Set an indicator layer for each sub-criteria. The indicators for the natural environment sub-criteria include air quality index, water quality index and vegetation coverage. The indicators for the human environment sub-criteria include richness of cultural resources and community friendliness. The indicators for the infrastructure sub-criteria include convenient transportation and comfortable accommodation. The indicators for the health and wellness facilities sub-criteria include coverage of medical care points and proportion of healthy dining restaurants. Step S1.4: Through expert voting and data analysis, remove indicators with a data missing rate higher than the preset threshold and retain the final indicator layer.

[0030] When establishing the hierarchical structure of the evaluation index system for health and wellness tourism resorts, it was divided into three layers: the objective layer, the criterion layer, and the indicator layer. The objective layer clearly defines the comprehensive evaluation of health and wellness tourism resorts, which is the core objective of the entire evaluation system. The criterion layer, based on literature analysis and expert interviews, is divided into four main criteria: environmental quality, service facilities, health and wellness activities, and management services. These criteria are the key dimensions for evaluating health and wellness tourism resorts. The indicator layer consists of specific evaluation indicators set for each sub-criterion. The selection of these indicators is based on data availability and expert review, ensuring the scientific rigor and practicality of the evaluation system. This hierarchical structure allows for a systematic assessment of various aspects of health and wellness tourism resorts, providing a foundation for subsequent weighting.

[0031] Specifically, the target layer is a comprehensive evaluation of health and wellness tourism resorts. This is the ultimate goal of the evaluation system, used to measure the overall health and wellness benefits of the resort. The criteria layer is divided into four main criteria: environmental quality, service facilities, health and wellness activities, and management services. The environmental quality criteria are further subdivided into two sub-criteria: natural environment and human environment. The natural environment sub-criteria include specific indicators such as air quality index, water quality index, and vegetation coverage, which reflect the quality of the resort's natural ecological environment. The human environment sub-criteria include indicators such as richness of cultural resources and community friendliness, which assess the resort's humanistic atmosphere and cultural characteristics.

[0032] The service facility criteria are further divided into two sub-criteria: infrastructure and health and wellness facilities. The infrastructure sub-criteria include indicators such as accessibility and accommodation comfort, while the health and wellness facilities sub-criteria include indicators such as the coverage of medical care points and the proportion of healthy dining restaurants. These indicators assess whether the resort's service facilities meet health and wellness needs. In setting the indicator layer, indicators with missing data rates exceeding a preset threshold were removed through expert voting and data analysis, ensuring the validity and reliability of the indicators.

[0033] Furthermore, when dividing the criteria and indicator layers, the specific content of each criterion and indicator can be further refined. For example, in the environmental quality criterion, the air quality index can be calculated by monitoring data such as pollutant concentrations and the number of days with good air quality within the resort area; the water quality index can be assessed by detecting indicators such as chemical oxygen demand and ammonia nitrogen content in water bodies; and vegetation coverage can be determined through satellite remote sensing data or field surveys. In the service facilities criterion, transportation convenience can be measured by assessing parameters such as the distance between the resort area and major transportation hubs and the convenience of public transportation; and accommodation comfort can be assessed by surveying tourists' satisfaction with accommodation facilities and the completeness of room facilities. In the health and wellness activities criterion, indicators such as the types of health and wellness activities, participation rates, and activity frequencies can be included, and these indicators can be obtained through questionnaires and activity records. In the management service criterion, indicators such as service quality, complaint handling efficiency, and safety management can be included, and these indicators can be assessed through tourist feedback and management records. Through these specific operational steps and parameter settings, the comprehensiveness and scientific nature of the evaluation indicator system can be ensured, providing a reliable basis for the evaluation of health and wellness tourism resorts.

[0034] In some embodiments, calculating the expert weights in step S2 specifically includes: Step S2.1: Using the proportional scaling method, each expert independently constructs a judgment matrix; Step S2.2: Calculate the expert confidence coefficient for each expert. This coefficient is based on industry experience value, professional background value, years of practice value, professional title level value, and historical evaluation accuracy value. The historical evaluation accuracy value is obtained by comparing the consistency between the expert's historical evaluation and the actual results. Step S2.3: Normalize the expert confidence coefficients to obtain the expert weight for each expert; Step S2.4: If any expert's expert weight is lower than the preset weight threshold, then exclude that expert and recalculate the expert weight; Step S2.5: Use expert weights to perform a weighted arithmetic average on multiple judgment matrices to obtain a comprehensive judgment matrix.

[0035] Specifically, the calculation of expert weights involves multiple parameters, including industry experience, professional background, years of experience, professional title, and historical evaluation accuracy. These parameters reflect an expert's actual work experience, professional knowledge level, length of service, professional title, and the accuracy of their historical evaluations in the health and wellness field. Industry experience can be measured by the number of years an expert has worked in the health and wellness industry; professional background can be assessed through the expert's education, professional field, and relevant certificates; years of experience directly reflect the expert's service time in the health and wellness industry; professional title can be quantified based on the expert's professional title level; and historical evaluation accuracy is determined by comparing the consistency between the expert's past evaluation results and the actual results. After normalization, these parameters yield the expert weight for each expert. If any expert's weight is lower than a preset weight threshold, that expert is excluded, and the weights of other experts are recalculated to ensure that all participating experts possess sufficient professional expertise and evaluation capabilities.

[0036] Furthermore, when calculating expert weights, industry experience can be comprehensively evaluated based on the number of years an expert has worked in the health and wellness industry and the number of projects they have participated in; professional background can be quantified by the expert's highest degree, the relevance of their professional field to the health and wellness industry, and the number of professional certificates they have obtained; years of experience can be directly calculated using the expert's service time in the health and wellness industry; professional title level can be quantified based on the expert's professional title level, such as assigning higher weight to senior titles; and historical evaluation accuracy can be determined by calculating the correlation coefficient between the expert's historical evaluations and actual results. During the normalization process, the value of each parameter can be converted to between 0 and 1 to ensure comparability between different parameters. This detailed calculation method can more accurately reflect the professional level and evaluation ability of each expert, thus providing a more scientific expert weight for the evaluation of health and wellness tourism resorts.

[0037] In some embodiments, the formula for calculating the expert confidence coefficient in step S2.2 is:

[0038] in, This represents the expert confidence coefficient of the j-th expert. This represents the normalized industry experience value. This represents the normalized professional background value. This represents the normalized value of years of service. This represents the normalized professional title grade value. This represents the normalized historical evaluation accuracy value. The weights are predefined weights, and satisfy the following conditions: Among them, the historical evaluation accuracy value The correlation coefficient was calculated by comparing historical evaluations by experts with real data.

[0039] Specifically, the calculation of the expert confidence coefficient involves multiple parameters. The industry experience value refers to the expert's work experience in the health and wellness industry and the number of projects they have participated in, which is normalized to a value between 0 and 1. The professional background value refers to the expert's education level, the relevance of their professional field to the health and wellness industry, and the number of professional certificates they have obtained, also normalized to a value between 0 and 1. The years of service value refers to the expert's service time in the health and wellness industry, expressed directly in years. The professional title level value refers to the expert's professional title level, assigned different values ​​according to the title's rank. The historical evaluation accuracy value refers to the consistency between the expert's past evaluation results and the actual results, obtained by calculating the correlation coefficient. After normalization, these parameters are weighted and summed using predefined weighting coefficients to obtain the confidence coefficient for each expert. The weighting coefficients are pre-set according to the importance of each parameter, summing to 1 to ensure a balance in the contributions of different parameters.

[0040] Furthermore, when calculating the expert confidence coefficient, the industry experience value can be comprehensively assessed by the expert's years of experience in the health and wellness industry and the number of projects they have participated in. This data can be obtained from the expert's resume and project records. The professional background value can be quantified by the expert's highest degree, the relevance of their professional field to the health and wellness industry, and the number of professional certificates they have obtained. This can be verified by the expert's academic and professional certificates. The years of experience value can be directly calculated using the expert's service time in the health and wellness industry, calculated as the difference between their entry date and current date. The professional title value can be quantified based on the expert's professional title level; for example, senior titles can be assigned higher weight, intermediate titles second, and junior titles the lowest. The historical evaluation accuracy value can be determined by calculating the correlation coefficient between the expert's historical evaluations and actual results. This can be calculated by comparing the expert's past evaluation records with actual results. Through these detailed calculation steps and parameter settings, the professional level and evaluation ability of each expert can be more accurately reflected, thus providing a more scientific expert weighting for the evaluation of health and wellness tourism resorts.

[0041] In some embodiments, step S3, calculating the preliminary weight vector and verifying its validity, specifically includes: Step S3.1: Calculate the largest eigenvalue and its corresponding eigenvector of the comprehensive judgment matrix using the eigenvector method; Step S3.2: Normalize the feature vectors to obtain the preliminary weight vectors; Step S3.3: Check if there are negative or zero values ​​in the initial weight vector. If so, return to step S2 to reconstruct the judgment matrix. Step S3.4: Calculate the variance of the initial weight vector. If the variance is lower than the preset variance threshold, then smooth the weights.

[0042] In calculating the initial weight vector and verifying its validity, the eigenvector method is used to calculate the largest eigenvalue of the comprehensive judgment matrix and its corresponding eigenvector. The eigenvector is then normalized to obtain the initial weight vector. This process ensures the normalization and consistency of the weight vector. Furthermore, this invention verifies the validity of the initial weight vector, including checking for negative or zero values ​​and calculating the variance to assess the distribution of the weights. These steps aim to ensure the rationality and reliability of the initial weight vector, providing a solid foundation for subsequent evaluation processes.

[0043] The eigenvector method is used to calculate the largest eigenvalue and its corresponding eigenvector of the comprehensive judgment matrix. The comprehensive judgment matrix is ​​constructed based on the pairwise comparison judgment matrices of experts, reflecting the relative importance of each evaluation indicator. The largest eigenvalue and its corresponding eigenvector can be obtained numerically. After normalization (each element divided by the sum of all elements), the eigenvector yields the initial weight vector. Normalization ensures that the sum of the weight vectors is 1, meeting the basic requirements for weights. Validity verification includes two main aspects: first, checking for negative or zero values ​​in the initial weight vector, as weights should be positive; negative or zero values ​​indicate an unreasonable aspect of the judgment matrix; second, calculating the variance of the initial weight vector to assess the distribution of weights. If the variance is lower than a preset variance threshold, it indicates that the weight distribution is too concentrated and requires smoothing.

[0044] When calculating the largest eigenvalue and eigenvector, numerical methods such as the power method or the Jacobian method can be used. These methods iteratively approximate the largest eigenvalue and its corresponding eigenvector. During the normalization process, it can be ensured that each weight value is between 0 and 1, and that the sum of all weight values ​​is 1. For validity verification, if negative or zero values ​​are found in the initial weight vector, the process can return to the expert judgment matrix construction step to reassess the reasonableness of the expert judgment. For variance calculation, a reasonable variance threshold can be set, such as 0.01. If the variance of the initial weight vector is lower than this threshold, the weight distribution can be considered too concentrated, requiring adjustment of the expert judgment matrix or recalculation of the weights to improve the weight distribution. Through these detailed steps and parameter settings, the calculation and verification process of the initial weight vector can be made more scientific and reliable, providing a more accurate weight basis for the evaluation of health and wellness tourism resorts.

[0045] In some embodiments, the qualified threshold for dynamically adjusting the consistency ratio in step S4 specifically includes: Step S4.1: Calculate the consistency index CI and consistency ratio CR of the comprehensive judgment matrix, where CI is calculated using the formula... Calculate, where, CR is the largest eigenvalue of the comprehensive judgment matrix, where n is the order of the comprehensive judgment matrix; CR is expressed by the formula... Calculate, where RI is the average random consistency index; Step S4.2: Set the initial pass threshold and adjust the threshold according to the importance of the indicator. The pass threshold for important indicators is lowered by 0.02 to 0.05, and the pass threshold for non-important indicators is raised by 0.01 to 0.03. Step S4.3: Analyze the data distribution of the index values, calculate the variance and skewness of each index, and if the variance is higher than the preset variance threshold and the skewness exceeds the preset skewness range, then further adjust the qualified threshold. Step S4.4: If the adjusted CR is still greater than the qualified threshold, return to step S2 to recalculate the expert weights or reconstruct the judgment matrix.

[0046] When dynamically adjusting the pass threshold for the consistency ratio, this invention considers the importance of the indicators and the data distribution. The consistency ratio is an important indicator for measuring the consistency of the judgment matrix, and it is obtained by calculating the ratio of the consistency index (CI) of the comprehensive judgment matrix to the average random consistency index (RI). This invention sets an initial pass threshold and adjusts the threshold according to the importance of the indicators, while simultaneously analyzing the data distribution of the indicator values ​​to dynamically adjust the pass threshold. This method can more scientifically reflect the consistency of each indicator and ensure the accuracy and reliability of the evaluation results.

[0047] Specifically, the calculation of the consistency ratio involves the consistency index CI and the average random consistency index RI. The consistency index CI is determined by the largest eigenvalue of the comprehensive judgment matrix. The formula is obtained by calculating the matrix order n. The average random consistency index (RI) is pre-calculated based on the order of the random matrix and is used to measure the consistency of the random matrix. The consistency ratio (CR) is the ratio of consistency (CI) to RI, i.e., CR = CI / RI. When dynamically adjusting the pass threshold, an initial pass threshold is first set, usually 0.1. Then, the threshold is adjusted according to the importance of the indicators: the pass threshold for important indicators is lowered by 0.02 to 0.05, and the pass threshold for unimportant indicators is raised by 0.01 to 0.03.

[0048] The data distribution of the indicator values ​​is analyzed, and the variance and skewness of each indicator are calculated. If the variance exceeds a preset variance threshold and the skewness exceeds a preset skewness range, the pass / fail threshold is further adjusted. These steps ensure that the pass / fail threshold for the consistency ratio can be dynamically adjusted according to the actual situation, improving the scientific rigor and reliability of the evaluation results.

[0049] When calculating the consistency index (CI), the maximum eigenvalue of the comprehensive judgment matrix can be calculated using numerical methods. Then, CI is calculated based on the matrix order n. When setting the initial pass threshold, a suitable value, such as 0.1, can be selected based on experience or experimental data. When adjusting the threshold according to the importance of the indicators, the importance of each indicator can be determined by expert ratings and the position of the indicator in the hierarchical structure. For example, experts can rate the indicators from 1 to 5 points, calculate the average score of each indicator, and combine the hierarchical depth of the indicator in the criterion layer and sub-criterion layer to calculate the importance score of the indicator. Based on the importance score, the indicators are divided into three categories: high importance, medium importance, and low importance, and different pass threshold adjustment ranges are set for each category. When analyzing the data distribution of indicator values, the variance and skewness of each indicator can be calculated. If the variance is higher than the preset variance threshold, such as 0.1, and the skewness exceeds the preset skewness range, such as -1 to 1, the pass threshold is further adjusted. Through these detailed steps and parameter settings, it can be ensured that the pass threshold of the consistency ratio can be dynamically adjusted according to the actual situation, improving the scientificity and reliability of the evaluation results.

[0050] In some embodiments, the importance classification of indicators in step S4.2 is based on expert ratings and the position of the indicators in the hierarchical structure, specifically including: Step S4.2.1: Experts score the indicators on a scale of 1 to 5, and calculate the average score for each indicator; Step S4.2.2: Calculate the importance score of the indicator by combining the hierarchical depth of the indicator in the criterion layer and the sub-criterion layer; Step S4.2.3: Based on the importance score, the indicators are divided into three categories: high importance, medium importance, and low importance, and different qualified threshold adjustment ranges are set for each category.

[0051] This invention, when adjusting the pass threshold for the consistency ratio, specifically considers the importance of the indicators and their position in the hierarchical structure. By having experts score the indicators and combining this with the hierarchical depth of the indicators in the criterion and sub-criterion layers, an importance score is calculated for each indicator. The indicators are then divided into three importance levels: high, medium, and low, with different pass threshold adjustment ranges set for each level. This method can more accurately reflect the importance of each indicator in the evaluation system, ensuring the scientific rigor and adaptability of the consistency test, thereby improving the accuracy and reliability of the entire evaluation system.

[0052] Specifically, the classification of indicator importance is based on two main factors: expert ratings and the indicator's position in the hierarchical structure. Expert ratings refer to experts assigning importance scores of 1 to 5 points to each indicator, with higher scores indicating greater importance. The average score for each indicator is obtained by averaging all expert ratings. The position of an indicator in the hierarchical structure refers to its depth within the criterion and sub-criterion layers. Deeper hierarchical depth generally indicates a more specific indicator and a more direct impact on the overall evaluation. Combining these two factors, an importance score is calculated for each indicator; higher scores indicate greater importance. Based on the importance scores, indicators are categorized into high, medium, and low importance. For high importance indicators, the pass / fail threshold is appropriately lowered, for example, by 0.02 to 0.05; for medium importance indicators, the pass / fail threshold remains unchanged; and for low importance indicators, the pass / fail threshold is appropriately raised, for example, by 0.01 to 0.03. This classification and adjustment method ensures that the pass / fail threshold for the consistency ratio matches the actual importance of the indicator.

[0053] Furthermore, in the expert scoring stage, multiple experienced experts can be invited to conduct independent scoring to reduce subjective bias. When calculating the importance score, different weights can be assigned to expert scores and hierarchy depth, for example, an expert score with a weight of 0.7 and a hierarchy depth with a weight of 0.3, and the final importance score is calculated using a weighted average. When adjusting the pass / fail threshold, the specific adjustment range can be flexibly set based on actual evaluation needs and empirical data. For example, for particularly critical high-importance indicators, the pass / fail threshold can be lowered to a lower level, such as 0.01, to ensure that the consistency of these indicators reaches a higher standard. Through these detailed operational steps, the pass / fail threshold for the consistency ratio can be adjusted more scientifically, further improving the scientific rigor and practicality of the evaluation system.

[0054] In some embodiments, the correction using the fuzzy comprehensive evaluation method (FCE) in step S5 specifically includes: Step S5.1: Determine the evaluation set as five levels, including excellent, good, average, poor, and very poor; Step S5.2: Collect expert evaluations of each indicator in the indicator layer and construct a fuzzy evaluation matrix R, where the elements... This represents the proportion of experts who gave the j-th level evaluation for the i-th indicator; Step S5.3: Perform multi-level fuzzy synthesis operation. First, perform primary synthesis on the indicators within the indicator layer, and then perform secondary synthesis on the criteria layer to obtain the fuzzy evaluation result vector B. Step S5.4: Defuzzify the fuzzy evaluation result vector B, and use the weighted average method to obtain the comprehensive score vector S for each indicator; Step S5.5: Iteratively correct the initial weight vector W using the comprehensive score vector S until the weight change is less than the preset convergence threshold.

[0055] The evaluation set is a collection of five levels: excellent, good, average, poor, and very poor, used to evaluate the indicators. The fuzzy evaluation matrix is ​​constructed based on expert evaluations of each indicator in the indicator layer, where each element represents the proportion of experts who gave a specific evaluation level for a particular indicator. The multi-level fuzzy synthesis operation consists of two stages: primary synthesis and secondary synthesis. Primary synthesis is performed on the indicators within each criterion, combining the preliminary weight vector with the fuzzy evaluation matrix to obtain the fuzzy evaluation result for each criterion. Secondary synthesis combines the primary synthesis result with the criterion layer weights to obtain the overall fuzzy evaluation result vector. Defuzzification transforms the fuzzy evaluation result vector into a specific comprehensive score vector, typically using a weighted average method. Finally, the preliminary weight vector is iteratively corrected using the comprehensive score vector until the weight change is less than a preset convergence threshold. These steps ensure the scientific and systematic nature of the weight correction process.

[0056] Furthermore, some operational details can be refined when implementing the above steps. For example, when constructing the fuzzy evaluation matrix, multiple experts can be invited to independently evaluate each indicator to ensure the diversity and objectivity of the evaluation. In the multi-level fuzzy synthesis operation, the initial synthesis can be achieved by multiplying the preliminary weight of each indicator with the corresponding fuzzy evaluation matrix to obtain the fuzzy evaluation result for each criterion. The second-level synthesis combines these fuzzy evaluation results with the criterion layer weights to obtain the overall fuzzy evaluation result vector. In the defuzzification process, a weighted average method can be used, multiplying each element in the fuzzy evaluation result vector by its corresponding weight and then summing them to obtain the comprehensive score for each indicator. During the iterative correction process, a small convergence threshold can be set, such as 0.001. When the weight change is less than this threshold, the weights are considered to have converged, and the iteration stops. Through these refined operational steps, the initial weight vector can be corrected more scientifically, further improving the scientificity and practicality of the evaluation system.

[0057] In some embodiments, the multi-level fuzzy synthesis operation in step S5.3 specifically includes: Step S5.3.1: For each criterion, perform initial synthesis using the preliminary weight vector W and the fuzzy evaluation matrix R, calculated as follows: ,in, This is the fuzzy evaluation result of the k-th criterion. It is the initial weight sub-vector corresponding to the k-th criterion. It is the fuzzy evaluation submatrix corresponding to the k-th criterion, with the symbol... Represents a fuzzy composition operator; Step S5.3.2: Convert the initial synthesis result As input for the second-level synthesis, the weights of the criterion layer are combined to perform the second-level synthesis, resulting in the overall fuzzy evaluation result vector B.

[0058] This invention employs a hierarchical processing method when performing multi-level fuzzy synthesis operations. This method first performs a primary synthesis on the indicators within each criterion, then uses the primary synthesis results as input for a secondary synthesis, ultimately obtaining the overall fuzzy evaluation result vector. This hierarchical processing approach can more meticulously reflect the relationships between various indicators and criteria, ensuring the scientific validity and accuracy of the fuzzy comprehensive evaluation method (FCE). In this way, expert fuzzy evaluation opinions can be effectively transformed into specific evaluation results, providing a basis for correcting the initial weight vector.

[0059] Specifically, the multi-level fuzzy synthesis operation includes two main steps: primary synthesis and secondary synthesis. Primary synthesis is performed on the indicators within each criterion, using a preliminary weight vector and a fuzzy evaluation matrix to obtain the fuzzy evaluation result for each criterion. Here, the preliminary weight vector represents the weight distribution of the indicators within each criterion, while the fuzzy evaluation matrix reflects the experts' evaluation opinions on these indicators. Secondary synthesis combines the fuzzy evaluation results obtained from primary synthesis with the criterion-level weights to obtain the overall fuzzy evaluation result vector. This process involves considering the criterion-level weights, which reflect the importance of different criteria in the entire evaluation system. Through this hierarchical synthesis method, it can be ensured that the evaluation opinions of each indicator and criterion are fully considered, and the final fuzzy evaluation result vector can more accurately reflect the overall evaluation situation.

[0060] Furthermore, when implementing multi-level fuzzy synthesis operations, some operational steps can be further refined. For example, in the initial synthesis stage, a separate fuzzy evaluation matrix can be constructed for each indicator within each criterion to ensure that the evaluation opinions for each indicator are accurately recorded. During the fuzzy synthesis operation, appropriate fuzzy synthesis operators can be used, such as the max-min synthesis operator or the weighted average synthesis operator; the specific choice depends on the characteristics of the evaluation indicators and the evaluation purpose. In the second-level synthesis stage, the weights of the criterion layer can be normalized to ensure that the sum of the weights is 1, thereby guaranteeing the rationality and comparability of the overall fuzzy evaluation result vector.

[0061] Furthermore, during the processing, sensitivity analysis can be performed on the fuzzy evaluation results to assess the impact of different parameter settings on the final result, thereby further optimizing the evaluation model. Through these refined operational steps, multi-level fuzzy synthesis operations can be performed more scientifically, improving the accuracy and reliability of the fuzzy comprehensive evaluation method.

[0062] In some embodiments, the iterative correction in step S5.5 specifically includes: Step S5.5.1: Set the initial weights to the initial weight vector W, and the iteration counter t=0; Step S5.5.2: Calculate the comprehensive score vector S under the current weight; Step S5.5.3: Use the formula Update the weights, where, It is the weight of the i-th indicator in the t-th iteration. It is the weight of the i-th indicator in the (t+1)-th iteration. It is the comprehensive score of the i-th indicator. It is the weight of the k-th indicator in the t-th iteration. It is the comprehensive score of the k-th indicator, and n is the total number of indicators; Step S5.5.4: Calculate the weight change ,in, It is the change in weight. It is the weight of the i-th indicator in the (t+1)-th iteration. It is the weight of the i-th indicator in the t-th iteration, where n is the total number of indicators; if If the number of iterations is less than the preset convergence threshold or the maximum number of iterations is reached, the iteration stops and the final weight is output; otherwise, t=t+1 and return to step S5.5.2.

[0063] This invention employs a dynamic adjustment method based on a comprehensive score vector when iteratively correcting the initial weight vector. This method calculates a comprehensive score for each indicator and uses these scores to iteratively update the initial weight vector until the weight change is less than a preset convergence threshold. This iterative correction process ensures the stability and accuracy of the final weights, allowing them to better reflect the importance of each indicator in the evaluation system.

[0064] Specifically, the iterative correction process includes setting initial weights as a preliminary weight vector and initializing an iteration counter. Then, a comprehensive score vector is calculated under the current weights, reflecting the overall performance of each indicator under that weight. Next, the weights are updated using the comprehensive score vector, adjusting the weights of each indicator according to a specific update rule. After the weights are updated, the weight change, i.e., the difference between the old and new weights, is calculated. If the weight change is less than a preset convergence threshold, or the maximum number of iterations is reached, the iteration stops, and the final weights are output. This process ensures the gradual optimization and convergence of the weights.

[0065] Furthermore, during the initialization phase, a preliminary weight vector can be used as the initial weights, and the iteration counter can be set to 0. When calculating the comprehensive score vector, a weighted average method can be used, multiplying the score of each indicator by its corresponding weight and then summing the results to obtain the comprehensive score. When updating the weights, a proportional update rule can be used: the new weight of each indicator equals its old weight multiplied by its comprehensive score, then divided by the sum of the products of the old weights and comprehensive scores of all indicators. This update rule can dynamically adjust the weights based on the comprehensive scores of the indicators. When calculating the change in weights, Euclidean distance or other suitable distance metrics can be used to calculate the difference between the old and new weights. Through these refined operational steps, iterative corrections can be performed more scientifically, ensuring the stability and accuracy of the final weights.

[0066] The above embodiments of the present invention have the following beneficial effects: The above description is merely an explanation of some preferred embodiments of the present invention and the technical principles employed. Those skilled in the art should understand that the scope of the invention as described in the embodiments of the present invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A method for determining the weight of an evaluation index of a health-care type tourism resort, characterized in that, The method comprises the following steps: Step S1: establishing a hierarchical structure of the evaluation index system of the health-care type tourist resort, the hierarchical structure comprising a target layer, a criterion layer and an index layer, wherein the criterion layer is divided into multiple criteria based on predefined rules, and the index layer determines specific indexes based on data availability and expert opinions; Step S2: based on the analytic hierarchy process (AHP), constructing multiple expert pairwise comparison judgment matrices for the evaluation indexes, introducing an expert confidence coefficient, calculating the expert weight of each expert based on the industry experience value, professional background value, service length value, title level value and historical evaluation accuracy value, and performing weighted average on the multiple judgment matrices based on the expert weight to obtain a comprehensive judgment matrix; Step S3: based on the comprehensive judgment matrix, calculating a preliminary weight vector of the evaluation indexes and verifying the validity of the preliminary weight vector; Step S4: performing consistency check on the comprehensive judgment matrix, calculating a consistency ratio, and dynamically adjusting a qualified threshold of the consistency ratio according to the index importance and data distribution, and through an iterative loop until the consistency ratio is less than or equal to the qualified threshold; Step S5: using the fuzzy comprehensive evaluation (FCE) method to modify the preliminary weight vector, including constructing a fuzzy evaluation matrix, determining an evaluation set, performing multi-level fuzzy synthesis operation, and adjusting the preliminary weight vector according to the synthesis result to obtain a final evaluation index weight.

2. The method of claim 1, wherein, The step S1 of establishing the hierarchical structure specifically comprises: Step S1.1: defining the target layer as the comprehensive evaluation of the health-care type tourist resort; Step S1.2: based on literature analysis and expert interviews, dividing the criterion layer into four criteria of environmental quality, service facilities, health-care activities and management services, wherein the environmental quality criterion comprises sub-criteria of natural environment and humanistic environment, and the service facilities criterion comprises sub-criteria of infrastructure and health-care special facilities; Step S1.3: setting an index layer for each sub-criterion, wherein the indexes of the natural environment sub-criterion include air quality index, water quality index and vegetation coverage, the indexes of the humanistic environment sub-criterion include cultural resource richness and community friendliness, the indexes of the infrastructure sub-criterion include traffic convenience and accommodation comfort, and the indexes of the health-care special facilities sub-criterion include medical care point coverage and healthy diet restaurant proportion; Step S1.4: through expert voting and data analysis, eliminating indexes with a data missing rate higher than a preset threshold, and retaining the final index layer.

3. The method of claim 1, wherein, The step S2 of calculating the expert weight specifically comprises: Step S2.1: using the proportion scale method to independently construct a judgment matrix by each expert; Step S2.2: calculating an expert confidence coefficient of each expert, the coefficient being based on the industry experience value, professional background value, service length value, title level value and historical evaluation accuracy value, wherein the historical evaluation accuracy value is obtained by comparing the consistency of the expert historical evaluation and actual results; Step S2.3: normalizing the expert confidence coefficient to obtain the expert weight of each expert; Step S2.4: if the expert weight of any expert is lower than a preset weight threshold, the expert is excluded and the expert weight is recalculated; Step S2.5: using the expert weight to perform weighted arithmetic average on the multiple judgment matrices to obtain the comprehensive judgment matrix.

4. The method of claim 3, wherein, The formula for calculating the expert confidence coefficient in step S2.2 is: wherein, denotes the expert confidence coefficient of the jth expert, denotes the normalized industry experience value, denotes the normalized professional background value, denotes the normalized working years value, denotes the normalized title rank value, denotes the normalized historical evaluation accuracy value, is a predefined weight coefficient, and satisfies wherein the historical evaluation accuracy value is calculated by the correlation coefficient of the expert historical evaluation and the true data.

5. The method of claim 1, wherein, The calculation of the preliminary weight vector in step S3 and the verification of validity specifically include: Step S3.1: Calculate the maximum eigenvalue of the comprehensive judgment matrix and its corresponding eigenvector using the eigenvector method; Step S3.2: Normalize the eigenvector to obtain the preliminary weight vector; Step S3.3: Check if there are negative or zero values in the preliminary weight vector. If there are, return to step S2 to rebuild the judgment matrix; Step S3.4: Calculate the variance of the preliminary weight vector. If the variance is lower than the preset variance threshold, smooth the weight.

6. The method of claim 1, wherein, The adjustment of the qualified threshold of the consistency ratio in step S4 specifically includes: Step S4.1: Calculate the consistency index CI and the consistency ratio CR of the comprehensive judgment matrix, is calculated, wherein, is the maximum eigenvalue of the comprehensive judgment matrix, and n is the order of the comprehensive judgment matrix. Compute, where Rl is the average random uniformity index; Step S4.2: Set the initial qualified threshold and adjust the threshold according to the importance classification of the index; Step S4.3: Analyze the data distribution of the index value, calculate the variance and skewness of each index, and if the variance is higher than the preset variance threshold and the skewness is beyond the preset skewness range, further adjust the qualified threshold; Step S4.4: If the adjusted CR is still greater than the qualified threshold, return to step S2 to recalculate the expert weight or rebuild the judgment matrix.

7. The method of claim 6, wherein, The importance classification of the index in step S4.2 is based on the expert score and the position of the index in the hierarchical structure, specifically including: Step S4.2.1: Score the index and calculate the average score of each index; Step S4.2.2: Calculate the importance score of the index based on the hierarchical depth of the index in the criterion layer and the sub-criterion layer; Step S4.2.3: According to the importance score, the index is divided into three categories: high importance, medium importance and low importance, and different qualified threshold adjustment amplitudes are set respectively.

8. The method of claim 1, wherein, The correction using the fuzzy comprehensive evaluation method FCE in step S5 specifically includes: Step S5.1: Determine the evaluation set as five levels, including excellent, good, general, poor and bad; Step S5.2: Collect the evaluation of each index in the index layer, and construct a fuzzy evaluation matrix R, where the elements represent the proportion of experts who give the ith index a jth level of evaluation. Step S5.3: Perform multi-level fuzzy synthesis operation. First, perform primary synthesis on the indexes in the index layer, then perform secondary synthesis on the criterion layer to obtain the fuzzy evaluation result vector B; Step S5.4: De-fuzzification processing of fuzzy evaluation result vector B, using weighted average method to obtain the comprehensive score vector S of each index; Step S5.5: Use the comprehensive score vector S to iteratively correct the preliminary weight vector W until the weight change is less than the preset convergence threshold.

9. The method of claim 8, wherein, The multi-level fuzzy synthesis operation in step S5.3 specifically includes: Step S5.3.1: For each criterion, use the preliminary weight vector W and the fuzzy evaluation matrix R to perform primary synthesis, the calculation formula is: wherein, is the fuzzy evaluation result of the kth criterion, is the preliminary weight sub-vector corresponding to the kth criterion, is the fuzzy evaluation sub-matrix corresponding to the kth criterion, the symbol denotes the fuzzy composition operator; Step S5.3.2: Combine the primary synthesis results As the input of the secondary synthesis, the secondary synthesis is combined with the criterion layer weights to obtain the overall fuzzy evaluation result vector B.

10. The method of claim 8, wherein, The iterative correction in step S5.5 specifically includes: Step S5.5.1: Set the initial weight as the preliminary weight vector W, and the iteration counter t=0; Step S5.5.2: Calculate the comprehensive score vector S under the current weight; Step S5.5.3: Using the formula updating the weights, wherein, is the weight of the i-th indicator at the t-th iteration, is the weight of the i-th indicator at the t+1-th iteration, is the overall score of the i-th indicator, is the weight of the k-th indicator at the t-th iteration, is the overall score of the k-th indicator, and n is the total number of indicators; Step S5.5.4: Calculate weight change amount wherein, is the weight change amount, is the weight of the i-th index at the t+1 iteration, is the weight of the i-th index at the t iteration, n is the total number of indexes; if is less than a preset convergence threshold or the number of iterations reaches a maximum number of iterations, the iteration is stopped and the final weight is output; otherwise, t=t+1, return to step S5.5.2.

Images