Clustering Models for Data Stream Mining

The scope of this research is to aggregate news contents that exists in data streams. A data stream may have several research issues. A user may only be interested in a subset of these research issues; there could be many different research issues from multiple streams, which discuss similar topic from different perspectives. A user may be interested in a topic but do not know how to collect all feeds related to this topic. The objective is to cluster all stories in the data streams into hierarchical structure for a better serve to the readers. The work utilizes segment wise distributional clustering that show the effectiveness of data streams. To better serve the news readers, advance data organization is highly desired. Once catching a glimpse of the topic, user can browse the returned hierarchy and find other stories/feeds talking about the same topic in the internet. The dynamically changing of stories needs to use the segment wise distributional clustering algorithm to have the capability to process information incrementally.