Incremental Clustering Algorithms
摘要
This chapter presents the general scheme for incremental clustering algorithms. These algorithms start by computing the center of the entire dataset and continue by adding one cluster center at each iteration. Two alternative procedures for finding initial cluster centers are given. Both of them utilize the incremental essence of clustering algorithms. Furthermore, we discuss the multi-start incremental clustering algorithm that uses one of these procedures. Finally, we describe the modified global k-means algorithm, fast-modified global k-means algorithm, and incremental k-medians algorithm. Detailed descriptions of all algorithms and their flowcharts are provided.