1.2 entropy weight method
According to the definition of information entropy, entropy value can be used to judge the dispersion degree of an index. The smaller the entropy value, the greater the dispersion degree of the index, and the greater the influence of the index on the comprehensive evaluation (that is, the weight). If all the values of an indicator are equal, the indicator will not play a role in the comprehensive evaluation.
Two calculation steps of entropy weight method
2. 1 Determine the index system
2.2 Data preprocessing
Data preprocessing refers to redundant data processing and outlier data processing.
2.3 standardized processing
Normalization means that indicators of the same dimension have different dimensions, and there are usually two methods:
1.0- 1 normalization (or critical value method), for example, the jj index of the second user is xijxij, and after normalization, it is X' Ijjj'. There are two formulas as follows:
x ' ij = xij？ Minimum (xj) and maximum (xj)? Min(xj) or x'ij=max(xj)? xijmax(xj)？ Minimum value (xj)
xij′= xij？ Minimum (xj) and maximum (xj)? Min(xj) or xij'=max(xj)? xijmax(xj)？ Minimum value (xj)
If the indicator is positive, select the first formula; If the indicator is auspicious, choose the second formula. Min(xj)min(xj) is the minimum value of jj index, and max(xj)max(xj) is the maximum value of jj index.
2.z score standardization
x ' ij = xij？ x？ Jet research …
xij′= xij？ x？ Jet research …
2.4 Calculate the entropy and weight of indicators
Firstly, calculate the ratio of the second user's jj index: yij = x' ij ∑ mi =1x' ijyij = xij' ∑ i =1mxij'
Calculate the information entropy EJ of jj index =? K∑mi= 1yijlnyijej=？ K∑i= 1myijlnyij, where KK is a constant and K= 1lnmK= 1lnm.
Calculate the weight of jj index wj= 1? ej∑j 1？ ejwj= 1？ ej∑j 1？ Electrojet perforation
2.5 Index weighted calculation score
Calculate the score or evaluation value S=∑j 100 by using the weighted summation formula? yijwjS=∑j 100？ Yijwj, where SS is the comprehensive score and wjwj is the first.