关键词: 面板数据聚类, 主成份分析, 动态时间规整

Abstract: Previous method on panel data clustering has some defects, so it is necessary to further improve and perfect panel data clustering method based on Dynamic Time Warping. This method is used to analysis the data of China national economic and technological development zone. The result of the empirical analysis shows that the new method not only reflects the dynamic change of panel data very well, but also avoids the problem of weighting the various distances in existing panel data clustering method. The clustering result is relatively stable and has a good visual effect.

Key words: Panel Data clustering, Principal Component Analysis, Dynamic Time Warping