Downloads: 130 | Views: 201

Research Paper | Computer Science & Engineering | India | Volume 6 Issue 1, January 2017

K Prototype Clustering with Efficient Summarization for Topic Evolutionary Tweet Stream Clustering

Swapnil Rajaram Ahire | Lakshita Landge

Abstract: Social media data reflects the lot of interests of virtual communities in a spontaneous and timely manner. Tweets are created as short text message and shared for each users and information analysts. Twitter that receives over four hundred million tweets per day has emerged as a useful supply of reports, blogs, opinions and additional. Our projected work consists of 3 parts. Tweet stream clustering to cluster tweets with k-prototype cluster algorithm (Existing paper shows, k-means clustering algorithm wont to produce the initial clusters. With globular clusters, it did not work well. Therefore in our projected work, we are using k-prototype algorithm to get tighter clusters than k-means algorithm, particularly if the clusters area globular). Second is tweet cluster vector technique to get rank summarization with greedy algorithm. Third to notice and monitors the outline summary based and volume based variation to provide timeline from tweet stream. Implementing continuous tweet stream reducing a text document is but not an easy task. Since a large variety of tweets area trashy, unrelated and raucous in nature, owing to the social nature of tweeting. Further, tweets area powerfully related with their announce instance and current tweets tend to reach an awfully quick rate. Efficiency - tweet streams are huge, thus the summarization algorithm ought to be greatly capable and time efficient. Flexibility - it ought to offer tweet summaries of random moment durations. Topic evolution - it ought to habitually notice sub - topic changes and therefore the moments that they happen.

Keywords: Tweet stream, summarization, timeline, summary, continuous tweet stream

Edition: Volume 6 Issue 1, January 2017,

Pages: 769 - 774