Regional Economic Indicators Analysis Based on Data Mining

Our research focuses on the problem that traditional economic can not provide exhaustive factors influencing the economy. By the study on data mining technologies, we try to find some methods adapting to regional economics. It is found through k-means and CADD algorithms that: high-dimension data is relatively sparse in space distribution and the distance is almost equal. So the algorithm based on distance and density has unideal clustering results when processing high-dimension data. The attributes related to clustering usually owns a few number of dimensions. The other number of dimensions will generate noise to affect the final result. Weighted CADD can partition the cluster based on adaptive density reachable ideas, which improve the clustering efficiency. Then according to the advantages of data mining technologies and the characteristics of regional economic data, we perform comparative analysis on 30 areas of China with two algorithms, from the natural attributes and economic attributes, to provide more scientific basis for coordinated development of Chinese regional economics.