An unsupervised clustering approach based on its data distribution and rectangle division

• Main Authors: Sheng-Hang Jong, 鐘晟航
• Other Authors: Ying-Kuei Yang
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/09219502747557882227

碩士 === 國立臺灣科技大學 === 電機工程系 === 96 === The main purpose of this paper is extended based on “An unsupervised clustering approach based on its data distribution” to offer a better way of cutting apart between the cluster and its neighbors. The new method presented in this paper will cut apart clusters by rectangle division. Comparing with round division, rectangle division can save some space and makes the clustered result more correct when dividing tall and slender data sets. The algorithm presented in this paper has been implemented, analysed and tested on six data sets. The results show that the proposed algorithm has much better classified ability than the Fuzzy c-Means algorithm and the algorithm of “An unsupervised clustering approach based on its data distribution”.