Classification and Evaluation for the Midwest Regional Innovation Capability Based on Principal Component Analysis and Self-organizing Neural Network

摘要

The imbalance of regional innovation capability is significantly prominent. Low regional innovation capacity of the Midwest limits the sustainable economic development in these regions. This paper selects main indexes data from high technology industrial technology activities in Midwest provinces in 2007 China's high technology industry statistics yearbook. Firstly, it reduces the correlations between variables by the principal component analysis. Secondly, it characterizes sample characteristic extracting 5 main components from 20 variables. Then it extracts 5 main components as input variables to build simulation model by using self-organizing neural networks. Midwest provinces are classified into seven groups. At last, it analyzes the classification reasons.