An Effective Compound Clustering Algorithm
Jianlin Zhang, Wensheng Zou, Jianfeng Xu, Lan Liu · 2009
There usually exit some reactive and redundant attributes in clustering objects. In order to improve the efficiency and veracity of clustering, we must delete those reactive and redundant attributes before clustering. A compound clustering algorithm is proposed in this paper. The algorithm first introduces fuzzy clustering to classify attributes, and then uses Fuzzy C-means (FCM) algorithm to partition objects and verify which attributes are redundant. The effectiveness of the proposed compound clustering algorithm is demonstrated with the Fisher Iris data set.