On smart Selection of Clustering Algorithms

Currently, a large number of clustering algorithms are available for data mining. But it will be difficult for people who to a large extent know little about data mining to select an appropriate clustering algorithm. In order to solve this problem, in this paper, we first comprehensively analyze a number of clustering algorithms, then summarize their evaluation criteria and apply the so-called fuzzy comprehensive evaluation to smart comprehensive evaluation for clustering algorithm. Finally, we propose a smart choice of specific data mining algorithm to help the users who lacks the corresponding expertise.