A cluster analysis method based on combination similarity and its application
Dan Wang, Ren Xiao-yu · 2010
With respect to the shortcomings that traditional cluster analysis method which only a single of similar measure portrait is utilized, the paper presents a new framework for combination similarity analysis. Investment clustering analysis have been applied for Shen Zhen listed companies. Moreover, the combination similarity analysis method is more comprehensive than the traditional single one, becasuse of the use of the position and shape to measure samples similarity. Clustering results will also be more effective and convincing. The new clustering method is intuitive, simple and easy to operate, can be widely used in pattern recognition and many other fields of economic management.