CN111062548A - Enterprise decision-making auxiliary real-time data acquisition and evaluation method - Google Patents

Abstract

The invention relates to an enterprise decision-making auxiliary real-time data acquisition and evaluation method, which is a decision-making auxiliary analysis system provided by enterprises, wherein an evaluation object and an evaluation index of the evaluation system are selected by a user according to self requirements, index weight is determined by combining a subjective (G2 weighting method) mode and an objective (CRITIC weighting method), comprehensive score of the evaluation object is automatically calculated by a computer through a TOPSIS method, and results are automatically output, so that efficient, scientific and reliable analysis results are provided for enterprise decision makers.

Description

Enterprise decision-making auxiliary real-time data acquisition and evaluation method

Technical Field

The invention belongs to the field of information, and particularly relates to an information data acquisition and evaluation method

Background

The invention provides a decision-making auxiliary analysis system for enterprises, an evaluation object and an evaluation index of the evaluation system are selected by a user according to self requirements, index weight is determined by combining a subjective (G2 weighting method) mode and an objective (CRITIC weighting method), comprehensive score of the evaluation object is automatically calculated by a computer through a TOPSIS method, results are automatically output, and efficient, scientific and reliable analysis results are provided for enterprise decision makers.

Disclosure of Invention

In order to help an enterprise solve the problems of difficult data acquisition and too subjective project evaluation in the project evaluation process, the invention provides an enterprise decision-making auxiliary real-time data acquisition and evaluation method, which specifically comprises the following steps:

an enterprise decision-making assisted real-time data acquisition method comprises the steps of manually selecting a target enterprise, selecting an evaluation index, inputting data, standardizing the data and collecting standardized data by a computer.

The data standardization refers to processing and processing the acquired data according to a fixed format to form a unified standard, so that data comparison is facilitated;

an enterprise decision-assisted evaluation method comprises the following steps:

(1) constructing an initial decision matrix A and a factor set U:

(2) data normalization process

(3) Weight determination

(4) Composite score calculation

Specifically, the above steps are determined according to the following method:

(1) constructing an initial decision matrix A and a factor set U:

firstly, a matrix with evaluation indexes of rows and evaluation objects of columns is constructed according to evaluation objects and an index system. Suppose that there are m evaluation objects x₁,x₂,x₃......x_m，X_i＝{x₁,x₂,x₃......x_mN evaluation indices, y₁,y₂,y₃……y_n，Y_j＝{y₁,y₂,y₃......y_n}, evaluation of object X_iIn the evaluation index Y_jIs a_ijThe evaluation value of which constitutes the initial decision matrix a ═ a_ij)_m×n(i＝1,2,…,m；j＝1,2,…,n)。

Assuming that the evaluation system has S first-level indexes, the index system is composed of S first-level indexes and j second-level indexes, and the total factor set U composed of the first-level indexes is equal to (U)₁,U₂,...,U_s) Sub-target factor set U formed by secondary indexes corresponding to the primary index k_k＝(U_k1,U_k2,…,U_km)(k＝1,2,…,s)。

(2) Data normalization process

Because the dimension and magnitude of the object represented by each index are different, the original data matrix needs to be calculated conveniently and quantitatively

Carrying out standardization treatment, and recording a standardized data matrix as T ═ T (T)_ij)_m×n(i-1, 2, …, m; j-1, 2, …, n), where t is_ijThe specific calculation formula is as follows:

wherein, for the index of positive effect type, the standardization processing formula (3) is adopted, and otherwise, the standardization processing formula (4) is adopted.

(3) Weight determination

Obtaining subjective weight of index by G2 method according to R_ikThe weight set of different obtained indexes with accurate assignment is W_ai＝(W_a1,W_a2,…W_as)。

And determining the objective weight of each index by using a CRITIC method, and importing the standardized index data into a system. Is a systemThe system calculates the difference and conflict of the selected indexes to further obtain the information content contained in the index j, the weight of the index j is correspondingly larger when the information content contained in the index is larger, and the objective index weight (W) is_bj) The calculation can be expressed as

The subjective weight obtained by expert evaluation and the objective weight obtained based on data calculation are consistent in importance, so the integrated weight is the average of the two weights, and the integrated weight of the index j can be expressed as

(4) Composite score calculation

Based on the integrated weight W_jAnd normalized data t_ijAnd calculating and obtaining Euclidean distance (D) between the evaluation object and the positive ideal solution and the negative ideal solution by using TOSISS⁺And D^-Equations 11 and 12), and thereby a composite score M (equation 13) of the evaluation target is obtained.

The method comprises the following steps: selecting n indexes x to be evaluated₁,x₂,x₃......x_nIn the evaluation index set { x }_i}＝{x₁,x₂,x₃......x_nSelecting the only one least important index as a reference object and marking as y_kRelabeling each index as y₁,y₂,y₃……y_nAnd n is the total index number. { x_i}＝{x₁,x₂,x₃......x_nAnd { y }_i}＝{y₁,y₂,y₃......y_nThere is a one-to-one correspondence. Will index y_iAnd the least important index y_kCompared to obtain the relative importance R_ikWherein R is_ik≥1，R_iiEach number represents the meaning as shown in table 1:

TABLE 1 relative importance between evaluation indices

1. Ratio of degree of importance R_ikFor point assignment scenarios

The domain expert determines the index y except the least important index according to the relevant information_kAs a unique reference, by applying the index y_iAnd the least important index y_kCompared to obtain the relative importance R_ikAnd further makes a rational judgment on the degree of importance of the evaluation index R_ikThe calculation method comprises the following steps:

if R is_ikThe assignment of (i) is accurate, then the weight coefficient W of the index i_iComprises the following steps:

2. ratio of degree of importance R_ikAssigning value cases to intervals

In some cases, the expert is on R_ikWhen subjective evaluation is performed, the assignment of R is not sure due to insufficient information_ikAn exact number, but with confidence, giving R_ikA range of values, but not disclaimed, i.e. R cannot be unambiguously assigned_ikTo which one and only one definite value is assigned, but which gives R with confidence_ikA G2 method with interval characteristics can be used. In this case, let R be the ratio of the degree of importance of the specified expert to the evaluation index based on the relevant information_ikGiving an interval D_ik。

Real bounded set of closed numbers d₁,d₂]＝(x|d₁≤x≤d₂X ∈ R) is called a closed interval, which can also be regarded as the closed interval from its end point d₁And d₂A pair of ordinal numbers, called interval numbers, is formed and is generally denoted by D. For D ═ D₁,d₂]Respectively, are called e (D) ═ d₂-d₁And n (d) ═ d₁+d₂) And/2 is the interval width and interval midpoint of D.

When n (D) is 0, D is a symmetric interval. For D₁＝[d₁₁,d₂₁]，D₂＝[d₁₂,d₂₂]Then, D is defined₁+D₂＝[d₁₁+d₁₂,d₂₁+d₂₂]. Often decisions are risky, called mapping

The method comprises the following steps of (1) performing an interval mapping function with expert risk attitude, wherein epsilon is a risk attitude factor, and the value range is as follows: -1/2 ≦ ε ≦ 1/2. For conservative experts, selecting-1/2 ≤ epsilon ≤ 0; for a neutral expert, taking epsilon as 0; for the risk type experts, 0-epsilon-1/2 is selected. For a given expert, ε is a known number.

Setting an evaluation index y of an expert according to related information_iAnd the least important index y_kRatio R of importance levels with respect to a criterion_ikGiving a number of intervals D_kI.e. to give R_ikThe value range of (a):

R_ik＝a_k∈[d_1k,d_2k]＝D_k,k＝1,2,…,m-1 (3)

wherein d is_1k≤d_2k,d_2m＝d_1m＝1

If { D_kThe assignment of the value is accurate, then

The second method comprises the following steps:

and objectively weighting the indexes by calculating two standards of difference and conflict between the evaluation indexes by using a CRITIC weighting method. The difference is obtained by taking the value difference of the same index between different samples through a standard deviation coefficient (S) on the basis of eliminating the influence of dimension_j) To make a measurement; between indexesThe conflict includes two aspects of size and direction, and is determined by the correlation coefficient (R)_pq) As shown, if the two indexes have a strong positive correlation, the conflict is low.

The standard deviation and correlation coefficient calculation mode of the jth index is as follows:

wherein p and q represent indices, R_pqA correlation coefficient sequence representing the index p and the index q

The conflict between the jth index and other indices may be quantified as

R_ijThe correlation coefficient between the j-th index and the i-th index is shown, and the collision between the indexes is the same for positive correlation and negative correlation with the same absolute value. Is provided with C_jThe information amount contained in the jth index is expressed as

Wherein C is_jThe larger the information amount contained in the jth index, the larger the weight of the j index, and the larger the information amount

The method 3 comprises the following steps:

and (3) by utilizing a TOPSIS method, calculating the distance between the index value and the positive ideal solution and the distance between the index value and the negative ideal solution, obtaining the relative closeness of each evaluation object and the optimal scheme, and sequencing, wherein the closer the evaluation object is to the ideal value, the higher the ranking is. The method is not limited by the sample amount, can obtain reasonable evaluation results through simple and convenient calculation, and is a multi-target decisionOne commonly used method in analysis. For data subjected to dimensionless normalization, the "positive ideal solution" is the maximum value in the normalized data to be evaluated, and the "negative ideal solution" corresponds to the minimum value in the normalized data to be evaluated. Positive ideal solution X⁺And negative ideal solution X^-Can be expressed as

The distance from the ideal value is calculated by using the weighted Euclidean distance, and the distances of the positive ideal solution and the negative ideal solution of the evaluation object are respectively expressed as D⁺And D^-：

D⁺And D^-Evaluation of the evaluation objects from different angles, D⁺The smaller the distance between the evaluation object and the optimal ideal solution is, namely the evaluation object is in a relative dominant position; d^-The larger the distance between the evaluation object and the worst ideal solution, the closer the evaluation object to the positive ideal solution, so D⁺And D^-The meaning of expression is to be considered consistent. Synthesis D⁺And D^-The closeness M between the evaluation object and the optimal plan can be obtained as the evaluation result of (2)^*That is, the composite score of the evaluation object:

the closeness M score calculated by the method is between 0 and 1, and the closer the score of the evaluation object is to 1, the smaller the distance between the evaluation object and the ideal solution is, and the better the evaluation object is compared with other evaluation objects.

Claims (3)

1. An enterprise decision-making assisted real-time data acquisition method comprises the steps of manually selecting a target enterprise, selecting an evaluation index, inputting data, standardizing the data and collecting standardized data by a computer.

2. The data acquisition method of claim 1, wherein the data standardization means that the acquired data are processed and processed according to a fixed format to form a unified specification, so as to facilitate data comparison.

3. An enterprise decision-assisted evaluation method comprises the following steps:

(1) constructing an initial decision matrix A and a factor set U:

(2) data normalization process

(3) Weight determination

(4) Composite score calculation

The data standardization means that the dimension and magnitude of the object represented by each index are different, so that quantitative calculation is facilitated, and an original data matrix is subjected to

And (6) carrying out standardization treatment.