A tweet grouping methodology utilizing inter and intra cosine similarity

Twitter enables users to write and publish messages with a maximum of 140 characters. This is sometimes termed micro-blogging because individuals often use Twitter to communicate their thoughts, commentary or feelings about any given subject. Twitter´s significant popularity and mass usage has resulted in any subject queried from the Twitter API that may return a vast number of tweets. These tweets can be related to several different categories. This paper proposes a hierarchical clustering system that groups tweets into meaningful clusters based on cosine similarity score.