Index screening method in analysis simulation evaluation process

Abstract

The invention discloses an index screening method in an analysis simulation evaluation process, and relates to the field of data processing in the simulation evaluation process. The method comprises the following steps: constructing a simulation evaluation data matrix, converting the simulation evaluation data matrix into an intuitionistic fuzzy matrix, converting a non-missing index value into an intuitionistic fuzzy index value, determining a reference index set, and filling missing values in the intuitionistic fuzzy matrix to complete data missing processing; according to the method, intuitive fuzzy processing is carried out on simulation data through the characteristic of the hesitation degree of the intuitive fuzzy set, then missing data is filled based on the intuitive fuzzy set algorithm, and finally missing simulation data is restored through the filled intuitive fuzzy number; the method not only can reasonably fill up data missing in simulation, but also eliminates unit and magnitude differences among simulation indexes, and is easy to implement.

Description

Technical Field

[0001] This invention relates to the field of analysis, simulation and evaluation technology, and in particular to a method for selecting indicators in the analysis, simulation and evaluation process. Background Technology

[0002] In analysis and simulation, the evaluation index system is a systematic method used to quantify, measure, and evaluate an entity or activity. It serves as a standard and method for quantifying and describing the object to be evaluated, and is the core and foundation of analysis and simulation evaluation. The evaluation index system plays a crucial role in the entire evaluation process. The establishment of the evaluation index system mainly involves selecting reasonable evaluation indicators from the original index system. The aim is to reduce the computational load and improve the evaluation efficiency without affecting the evaluation effect. However, in the evaluation process, it is often encountered that the number of indicators for the evaluation purpose and content is very large, leading to a decrease in evaluation efficiency. In particular, indicators with low importance not only have a negligible role in the evaluation results, but also greatly increase the computational load of the evaluation, reducing evaluation efficiency. In some cases, the excessive amount of evaluation index data can even cause the evaluation system to become sluggish and lead to the emergence of "digital black holes." Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for selecting indicators in the analysis and simulation evaluation process, which solves the problem of slow evaluation system and reduced evaluation efficiency caused by the excessive amount of evaluation indicator data.

[0004] The technical solution adopted in this invention is:

[0005] A method for selecting indicators in the analysis and simulation evaluation process, specifically including the following steps:

[0006] S1. Set evaluation values ​​and construct the evaluation matrix:

[0007] Suppose there are m samples and n indicators in an evaluation, and the number of indicators selected is t (t≤n). Construct the evaluation matrix. , Let represent the value of the j-th attribute index in the i-th sample, and determine the number of sample indicators selected as t, where t≤n;

[0008] S2. Type of sample values ​​for the judgment indicator:

[0009] The indicator sample values ​​are of three types: real number type, fuzzy number type, and intuitionistic fuzzy number type. The importance C of each indicator is calculated using different calculation methods for the three different types of sample values. j According to importance C j The size of the index is used to rank the indicators in the sample, and the top t evaluation indicators with the highest importance are selected.

[0010] The indicator selection method in the analysis, simulation, and evaluation process, where the sample value of the indicator is a real number, has an indicator importance C. j The solution method is as follows:

[0011] (1) Evaluation matrix H ( Normalization, to obtain the normalized matrix. ,

[0012] (1)

[0013] This represents the j-th index value in the i-th sample;

[0014] (2) Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F of matrix A:

[0015] ,in (2);

[0016] ,in (3);

[0017] (3) Calculate the importance C of each real-valued indicator. j :

[0018] (4).

[0019] The indicator selection method in the analysis, simulation, and evaluation process, where the sample values ​​of the indicators are fuzzy numbers, determines the indicator importance C. j The solution method is as follows:

[0020] make ,and , , ;

[0021] (1) Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F for each index in the fuzzy numerical model:

[0022] ,in (5);

[0023] ,in (6);

[0024] (2) Calculate the importance C of each index in the fuzzy numerical model. j :

[0025] (7).

[0026] The indicator selection method in the analysis, simulation, and evaluation process, where the sample values ​​of the indicators are intuitionistic fuzzy numbers, determines the indicator importance C. j The solution method is as follows:

[0027] make ,and , ;

[0028] (1) Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F for each index of the intuitionistic fuzzy number model:

[0029] ,in (8);

[0030] ,in (9);

[0031] (2) Calculate the importance C of each index in the intuitionistic fuzzy number model. j :

[0032] (10).

[0033] Due to the adoption of the technical solution described above, the present invention has the following advantages:

[0034] This invention analyzes the index selection method in the simulation evaluation process. By setting evaluation values ​​and constructing an evaluation matrix, the positive ideal value set Z and negative ideal value set F of each index are obtained according to the algorithms corresponding to three sample value types: real number type, fuzzy number type, and intuitionistic fuzzy number type. Thus, the importance C is calculated. j The invention sorts and simplifies the selection of evaluation indicators by calculating the importance of each indicator in the sample, eliminating indicators with insignificant discrimination and small impact on the evaluation results. This reduces the number of indicators in the evaluation process, improves the efficiency of evaluation calculation, and avoids the occurrence of "digital black holes," which is of vital importance to the efficiency of evaluation. Detailed Implementation

[0035] The present invention will be further explained and illustrated below with reference to embodiments. However, this should not be construed as limiting the scope of protection of the present invention. The purpose of disclosing the present invention is to protect all technical improvements within the scope of the present invention.

[0036] The index selection method in the analysis and simulation evaluation process described in this invention specifically includes the following steps:

[0037] 1. Set evaluation values ​​and construct an evaluation matrix.

[0038] Suppose there are m samples and n indicators in an evaluation, and the number of indicators selected is t (t≤n). Construct the evaluation matrix. , This represents the value of the j-th attribute index in the i-th sample;

[0039] 2. Determine the type of sample values ​​for the indicator.

[0040] The sample values ​​of the indicators are of three types: real number, fuzzy number, and intuitionistic fuzzy number. If the sample value of the indicator is real number, the importance of each indicator is calculated using Algorithm 1. If the sample value of the indicator is fuzzy number, the importance of each indicator is calculated using Algorithm 2. If the sample value of the indicator is intuitionistic fuzzy number, the importance of each indicator is calculated using Algorithm 3.

[0041] 3. Obtain the importance C of each indicator using Algorithm 1, Algorithm 2, or Algorithm 3. j

[0042] When the sample values ​​of the indicators are real numbers, this invention uses Algorithm 1 to calculate the importance of each indicator. The specific steps are as follows:

[0043] Step (1): Evaluate the evaluation matrix H ( Normalization, to obtain the normalized matrix. ;

[0044] (1);

[0045] in, This represents the j-th index value in the i-th sample;

[0046] Step (2): Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F of matrix A;

[0047] ,in (2);

[0048] ,in (3);

[0049] Step (3): Calculate the importance C of each real-valued indicator. j ,

[0050] (4);

[0051] When the sample values ​​of the indicators are fuzzy numbers, this invention uses Algorithm 2 to calculate the importance of each indicator. The specific steps are as follows: Let for ,and , , ;

[0052] Step (1): Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F for each index of the fuzzy numerical model;

[0053] ,in (5);

[0054] ,in (6);

[0055] Step (2): Calculate the importance C of each index in the fuzzy numerical model. j ,

[0056] (7);

[0057] When the sample values ​​of the indicators are intuitive fuzzy numbers, this invention uses Algorithm 3 to calculate the importance of each indicator. The specific steps are as follows: Let for ,and , ;

[0058] Step (1): Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F for each index of the intuitionistic fuzzy number model;

[0059] ,in (8);

[0060] ,in (9);

[0061] Step (2): Calculate the importance C of each index in the intuitionistic fuzzy number model. j ,

[0062] (10);

[0063] 4. According to importance C j The indicators are ranked according to their magnitude, and the top t indicators with the highest importance are selected. The value of t is generally determined based on the number of evaluation indicators and the specific evaluation object. The above steps complete the indicator screening in the evaluation process.

[0064] Example 1

[0065] In a certain evaluation, there are 10 samples and 8 indicators. The number of indicators selected is 4. The sample data are real numbers, and the indicator values ​​of each real number sample are shown in Table 1.

[0066] Table 1 Real Number Evaluation Data Table

[0067] sample Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6 Indicator 7 Indicator 8 1 3.00 0.65 29.00 95.00 55.00 20.20 17.00 0.01 2 5.00 0.87 31.00 94.00 60.00 25.70 12.00 0.04 3 4.00 0.41 42.00 87.00 56.00 22.10 19.00 0.05 4 7.00 0.45 37.00 82.00 59.00 27.30 11.00 0.09 5 3.00 0.68 43.00 93.00 57.00 26.80 14.00 0.03 6 4.00 0.72 46.00 96.00 56.00 24.40 16.00 0.06 7 5.00 0.37 39.00 89.00 56.00 25.70 19.00 0.01 8 5.00 0.56 45.00 99.00 56.00 29.10 18.00 0.02 9 6.00 0.75 39.00 82.00 58.00 28.70 12.00 0.09 10 5.00 0.61 49.00 95.00 55.00 23.80 15.00 0.07

[0068] (1) Construct the evaluation matrix H based on Table 1:

[0069]

[0070] (2) Since the sample values ​​are real numbers, the evaluation indicators are selected and the importance of the evaluation indicators is analyzed according to Algorithm 1;

[0071] (3) Calculate the normalized matrix A according to formula (1) in Algorithm 1:

[0072]

[0073] (4) Calculate the set of positive ideal values ​​Z according to formula (2) in Algorithm 1, and calculate the set of negative ideal values ​​F according to formula (3):

[0074] ;

[0075] The importance C of each real-valued indicator is calculated according to formula (4) in Algorithm 1. j (i.e., the set of C values)

[0076]

[0077] (5) Based on the importance C, the real number indicators can be sorted as follows: indicator 8 > indicator 2 > indicator 1 > indicator 7 > indicator 3 > indicator 6 > indicator 4 > indicator 5; therefore, the selected indicators are: indicator 8, indicator 2, indicator 1, and indicator 7.

[0078] Example 2:

[0079] In a certain evaluation, there are 7 samples and 7 indicators. The number of indicators to be selected is 4. The sample data is fuzzy, and the indicator values ​​of each fuzzy sample are shown in Table 2.

[0080] Table 2. Fuzzy Numerical Evaluation Data Table

[0081] sample Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6 Indicator 7 1 (0.50,0.50) (0.80,0.20) (0.75,0.25) (0.50,0.50) (0.68,0.32) (0.80,0.20) (0.89,0.11) 2 (0.60,0.40) (0.70,0.30) (0.85,0.15) (0.70,0.30) (0.75,0.25) (0.60,0.40) (0.91,0.09) 3 (0.70,0.30) (0.90,0.10) (0.86,0.14) (0.80,0.20) (0.73,0.27) (0.70,0.30) (0.82,0.18) 4 (0.60,0.40) (0.80,0.20) (0.84,0.16) (0.60,0.40) (0.72,0.28) (0.70,0.30) (0.94,0.06) 5 (0.70,0.30) (0.80,0.20) (0.82,0.18) (0.70,0.30) (0.70,0.30) (0.60,0.40) (0.94,0.06) 6 (0.60,0.40) (0.90,0.10) (0.80,0.20) (0.80,0.20) (0.65,0.35) (0.80,0.20) (0.86,0.14) 7 (0.80,0.20) (0.80,0.20) (0.79,0.21) (0.60,0.40) (0.67,0.33) (0.70,0.30) (0.93,0.07)

[0082] (1) Construct the evaluation matrix H based on Table 2:

[0083] (2) Since the sample values ​​are fuzzy numbers, the evaluation index selection and evaluation index importance analysis are performed according to Algorithm 2;

[0084] (3) Calculate the set of positive ideal values ​​Z according to formula (5) in Algorithm 2, and calculate the set of negative ideal values ​​F according to formula (6):

[0085] ;

[0086] The importance C of each index in the fuzzy numerical model is calculated according to formula (7) in Algorithm 2. j (i.e., the set of C values):

[0087]

[0088] (4) The importance C can be used to sort the fuzzy numerical indicators as follows: indicator 4 > indicator 1 > indicator 6 > indicator 2 > indicator 7 > indicator 3 > indicator 5; therefore, the selected indicators are: indicator 4, indicator 1, indicator 6, and indicator 2.

[0089] Example 3:

[0090] In a certain evaluation, there are 6 samples and 6 indicators. The number of indicators to be selected is 4. The sample data is intuitionistic fuzzy number, and the indicator values ​​of each intuitionistic fuzzy number sample are shown in Table 3.

[0091] Table 3. Evaluation Data Table of Intuitive Fuzzy Number Models

[0092] sample Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6 1 (0.87,0.11) (0.76,0.21) (0.70,0.20) (0.67,0.31) (0.89,0.06) (0.75,0.21) 2 (0.84,0.08) (0.75,0.13) (0.60,0.30) (0.64,0.30) (0.90,0.05) (0.78,0.19) 3 (0.87,0.07) (0.70,0.14) (0.80,0.10) (0.61,0.29) (0.86,0.09) (0.79,0.17) 4 (0.89,0.09) (0.79,0.09) (0.80,0.10) (0.68,0.24) (0.87,0.08) (0.74,0.18) 5 (0.86,0.08) (0.76,0.14) (0.70,0.20) (0.65,0.34) (0.94,0.01) (0.73,0.14) 6 (0.81,0.17) (0.75,0.18) (0.80,0.10) (0.62,0.28) (0.92,0.04) (0.78,0.21)

[0093] (1) Construct the evaluation matrix H based on Table 3:

[0094]

[0095] (2) Since the sample values ​​are intuitive fuzzy numbers, the evaluation index selection and evaluation index importance analysis are performed according to Algorithm 3;

[0096] (3) Calculate the set of positive ideal values ​​Z according to formula (8) in Algorithm 3, and calculate the set of negative ideal values ​​F according to formula (9):

[0097] ;

[0098] The importance C of each index in the intuitionistic fuzzy number model is calculated according to formula (10) in Algorithm 3. j (i.e., the set of C values):

[0099]

[0100] (4) The importance C can be used to rank the intuitionistic fuzzy number type indicators as follows: indicator 5 > indicator 1 > indicator 2 > indicator 6 > indicator 3 > indicator 4; therefore, the selected indicators are: indicator 5, indicator 1, indicator 2, and indicator 6.

[0101] This invention fully considers the sample value type and sorts and simplifies the evaluation indicators by calculating the importance of each indicator. When evaluating multiple sample indicators, the importance of the evaluation indicators can be compared according to different sample value types using Algorithm 1, Algorithm 2 and Algorithm 3. Based on the evaluation criteria of each sample indicator, the simplified evaluation indicators are further selected, thereby reducing the number of indicators in the evaluation process and improving the evaluation calculation efficiency.

[0102] The parts of this invention not described in detail are prior art.

[0103] The embodiments selected herein for the purpose of disclosing the inventive objectives are currently considered suitable; however, it should be understood that the invention is intended to include all variations and modifications of the embodiments that fall within the scope of this concept and invention.

Claims

1. A method for selecting indicators in the analysis and simulation evaluation process; characterized in that: Specifically, the following steps are included: S1. Set evaluation values ​​and construct the evaluation matrix: Suppose there are m samples and n indicators in an evaluation, and the number of indicators selected is t (t≤n). Construct the evaluation matrix. , Let represent the value of the j-th attribute index in the i-th sample, and determine the number of sample indicators selected as t, where t≤n; S2. Type of sample values ​​for the judgment indicator: The indicator sample values ​​are of three types: real number, fuzzy number, and intuitionistic fuzzy number. The importance C of each indicator is calculated using different calculation methods for each of the three different types of sample values. j According to importance C j Based on the magnitude of the index, the indicators in the sample are ranked, and the top t evaluation indicators with the highest importance are selected.

2. The index selection method in the analysis and simulation evaluation process according to claim 1, characterized in that: When the sample values ​​of an indicator are real numbers, the indicator importance C j The solution method is as follows: (1) Evaluation matrix H ( Normalization, to obtain the normalized matrix. , ；（1） This represents the j-th index value in the i-th sample; (2) Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F of matrix A: ,in (2); ,in (3); (3) Calculate the importance C of each real-valued indicator. j : （4）。 3. The index selection method in the analysis and simulation evaluation process according to claim 1, characterized in that: When the sample values ​​of an indicator are fuzzy numbers, the indicator importance C j The solution method is as follows: make ,and , , ; (1) Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F for each index in the fuzzy numerical model: ,in (5); ,in (6); (2) Calculate the importance C of each index in the fuzzy numerical model. j : （7）。 4. The index selection method in the analysis and simulation evaluation process according to claim 1, characterized in that: When the sample values ​​of an indicator are of the intuitionistic fuzzy number type, the indicator importance C j The solution method is as follows: make ,and , ; (1) Calculate the set of positive ideal values ​​Z and the set of negative ideal values ​​F for each index of the intuitionistic fuzzy number model: ,in (8); ,in (9); (2) Calculate the importance C of each index in the intuitionistic fuzzy number model. j : （10）。

Images