Evaluation of Systems Current Status by PCA-RBF Neural Network and Novel Fuzzy Intelligence Method

Abstract

Abstract The health and the sustainability of a higher education system are crucial to the prosperity of a nation. It is crucial to effectively evaluate a higher education system in various aspects and promptly adjust the corresponding educational policies. Firstly, 23 higher education quality indicators with respect to more than 1,000 universities are carefully collected and converted to the corresponding indicators of 40 countries/regions. Moreover, such indicators are normalized by the range transformation method in order to facilitate subsequent analyses. Secondly, a novel fuzzy intelligence (F&I) method is proposed to model the health states of different higher education systems based on radial basis function (RBF) neural networks. This method firstly reduces the input indicators’ dimension by the principal component analysis (PCA) technique. Using the PCA results as the input layer, the RBF neural network is carefully trained and the F&I score of each country/region is therefore obtained. Next, the hierarchical cluster analysis is carried out to depict the health state of each county/region. 1, 5, 5 and 29 countries/regions are categorized as healthy, good, general and unhealthy, respectively.