A Study on the CS Clustering Algorithm and its Applications

• Main Authors: Fu-Chien Cheng, 鄭富謙
• Other Authors: Yin-Te Tsai
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/32610591575276067660

碩士 === 靜宜大學 === 資訊管理學系 === 89 === More and more researchers apply the data mining technologies on the network which is a huge web database recently. The main goal of data mining is to find implied and useful information from very large databases. And now Web Mining is also rising and flourishing day by day with WWW. Web Usage Mining is the more important subject recently, and it’s focus is on analyzing web users’ browser behaviors — called User Session Patterns. We can analyze user session patterns to mine useful information implied on the network, and make Internet Service Providers (ISP), network flow managers and common users get benefits from it. Web usage mining can let us more know web users or customers, and integrate user groups. Compact Set (CS) clustering algorithm is a hierarchical clustering algorithm. Based upon the definition of our algorithm can find more cohesion clusters. In this paper, we also use CS clustering algorithm to analyze users’ behaviors on the web proxy log. From our experiments, the performance of clustering algorithm is better than other package software and its usage is more flexible. We believe that this algorithm is good for clustering study in data mining research.