Social Information Sub-topic Clustering Using News Tag

Discussions in social networks change frequently as users express their different points of view. Most of the topic tracking systems designed to follow these changes use a set of keywords to filter articles in order to identify key points or user opinions. These systems group all of the articles together even though some are closely related or provide no new information. Therefore, it is difficult to ascertain how many distinct topics of discussion have occurred in a given time period without reading all the articles. This research describes a method for dividing a tracked topic that contains a set of articles into several sub-topics automatically. We provide data scientists with a dashboard that directly displays sub-topic categories and allows them to identify important related issues over a daily or weekly period.