A Novel Clustering Algorithm Based on Self-Organization Procedure

• Main Author: 陳玉玲
• Other Authors: 洪文良
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/35045936095034037223

碩士 === 國立新竹教育大學 === 人資處數學教育碩士班 === 99 === Abstract This study presents a method to select the parameter in Chen and Shiu's (2007) clustering algorithm. The data points in the proposed clustering algorithm can self-organize local optimal cluster number without using cluster validity functions. The proposed clustering method is also robust to outliers based on the numerical experiments. Therefore, the proposed algorithms exhibits two robust clustering characteristics: (i) robust to the initialization (cluster number and initial guesses), (ii) robust to noise and outliers. Several numerical data and actual data sets are used in the proposed algorithm to show these good aspects.