Incremental clustering algorithm based on representative points
一种基于代表点的增量聚类算法

As the existing incremental clustering algorithms have various disadvantages such as high sensitivity to parameters, high time-space complexity, etc. This paper presented an incremental algorithm based on representative points. It first used the static clustering algorithm based on representative points to cluster the original data set. Then according the relationship between the new points and the existing representative points, the algorithm judged whether the new points should be added to the clusters containing the existing representative points or promoted as new representative points. Finally it used the static clustering algorithm again to cluster the new points. Experimental result shows that this algorithm is insensible to parameters, efficient and occupies little space.