Research on Multi-index Grey Clustering Approach Based on Adaptive Weight for Panel Data

摘要

In order to fully exploit the difference information between the clustering objects and object's spatiotemporal characteristic attributes, while considering the multiple information of panel data, we propose a grey clustering method of object-oriented spatiotemporal characteristics differences on adaptive weight by using the grey target decision-making method from the three dimensions of the absolute amount level, the increasing amount level, and the volatility level. The proposed method not only can degenerate into the traditional single attribute clustering algorithm but can also conduct clustering prediction for the future category of panel data. Finally, we use the new method has been applied for dealing with the clustering problems on the performance assessment of R&D activities from colleges and universities. The results show that the proposed approach can make full use of the spatiotemporal characteristic attributes' spatiotemporal difference information of panel data, and distinguish the existing problems of the clustering objects under different combinations of the spatiotemporal characteristic attribute, which can help obtain the more objective and practical clustering results.