| Popis: |
The principal component analysis theory is quite mature in the simplification and comprehensive evaluation of real data, and the traditional principal component analysis only strives for the form that the index value is real or interval number. It lacks certain persuasion in the face of large uncertainty index values, but it lacks the simplification of the three-parameter interval number. Based on the principal component analysis theory that the index data is an interval number, the paper proposed the M-PCA (three-parameter interval principal component analysis) method to screen out the unrelated indicators. Principal component analysis does not consider the influence of the negative factor when it is used to reduce the dimension of the comprehensive index. There are some misunderstandings and defects in the application. In dealing with the weights of unrelated indicators, this paper considered the subjective theories to solve the weights of the new indicators, combined with the multi-attribute gray target decision model to evaluate quality indicators. Finally, it combined with examples and verified that the method is more scientific, reasonable, and effective. |
