Designing China´s Local Government Performance Evaluation Indicators Based on Data Mining: An Exploratory Study in Four Municipal Governments in Jiangsu

In a tremendously large country like China, there are substantial differences on natural geographic features, history backgrounds, and social realities between the varied regions. Because of the heterogeneity, the local government performance evaluation indicator designing is a systematical engineering which could be affected by various factors. In order to solve the problems, we established a series of local government-oriented performance evaluation indicators database, data marts and data warehouses. On the basis of the data warehouse, this study uses RBF neural network as the soft clustering tool to excavate 10 stair indicators, 30 secondary indicators and 90 triple indicators under some support degrees.