On Summarization and Timeline Generation for Evolutionary Tweet Streams

Abstract

Short-instant messages, for example, tweets are being made and shared at an extraordinary rate. Tweets, in their crude shape, while being informative, can likewise be overwhelming. For both end-clients and information investigators, it is a bad dream to push through a huge number of tweets which contain tremendous measure of clamor and excess. In this paper, we propose a novel continuous synopsis system called Sumblr to reduce the issue. As opposed to the conventional report synopsis techniques which center around static what's more, little scale informational collection, Sumblr is intended to manage dynamic, quick arriving, and extensive scale tweet streams.