Skip to Main Content
To solve the inefficiency problem of traditional PAM algorithm for initial cluster center, a new improved PAM algorithm is proposed. The algorithm builds minimum spanning tree and then splits it to get k initial clusters with the relevant cluster centers. Theory analysis and experimental results show that the found initial centers are closed to the desired cluster centers, and the improved algorithm can achieve stable clustering results and higher cluster clustering accuracy.
Date of Conference: 22-24 June 2012