手机知网 App
24小时专家级知识服务
打 开
手机知网|搜索

HNBC:Hierarchical Neighborhood-based Clustering for Rare Events

HUANG Joshua Z

  A rare event in a data set is presented as a small cluster that is buried in other clusters.Rare events are very important in real world applications,for instance a small group of customers for a valuable niche market.In this paper,we propose a new clustering algorithm HNBC for rare event discovery.HNBC is a hierarchical version of the neighborhood-based clustering algorithm(NBC) that clusters data according to the local density which has advantages over the global density as used by DBSCAN.Experiments on synthetic data have shown that HNBC had a good performance in identifying rare events.……   
App内打开