Improved AHP and Manifold Learning Model for R&D And Transformation Functional Platform Performance Evaluation

Abstract

The accurate assessment of R&D and transformation functional platform performance is an important basis for the improvement of high-tech industry competitiveness. In this paper, the authors establish an improved AHP-manifold learning model to solve the problems of the traditional AHP method which needs to satisfy the consistency condition in constructing judgment matrices. In the ranking process of inconsistency of judgment matrices, on the basis of the neighbour distance, the neighbour distance matrices of the data sets corresponding to judgment matrices are constructed first. Next, each data point is mapped to a low-dimensional global coordinate system based on the linear representations of the neighbour points, and the low-dimensional embeddings corresponding to the judgment matrices are obtained. Then the ranking conclusion is obtained by analysing the superiority and inferiority ranking of the elements according to the correspondingly calculated low-dimensional embeddings from each hierarchy. Finally, the proposed method and another numerical method are used to assess R&D and transformation functional platform performance. The result illustrates that the proposed method has a higher level of effectiveness and practicability, and it can provide good guidance for improving platform performance.(original abstract)