A hybrid incremental clustering method—combining SVM and enhanced CBC algorithm

• Main Authors: Hsieh,Kong-Ling, 謝岡陵
• Other Authors: Chiu,Deng-Yiv
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/82273717239496436297

碩士 === 中華大學 === 資訊管理學系 === 94 === In the study, a new hybrid incremental clustering method is proposed in combination with SVM and enhanced CBC algorithm. SVM classifies the incoming document to see if it belongs to the existing classes. Then enhanced CBC algorithm is used to cluster the unclassified documents. In the algorithm, SVM can significantly reduce the amount of calculation and the noise of clustering. Enhanced CBC algorithm can effectively control the number of clusters, raise performance and the number of classes grows gradually based on the structure of current classes without clustering all of documents again. In experimental results, the hybrid incremental clustering outperforms the enhanced CBC clustering and other algorithms. Also, enhanced CBC clustering outperforms original CBC.