Feature-Rich Segment-Based News Event Detection on Twitter

Abstract

Event detection on Twitter is an important and challenging research topic. On the one hand, Twitter provides first-hand information and fast broadcasting. On the other, challenges include short and noisy content, big volume data and fast-changing topics. Dominant approaches for Twitter event detection model events by clustering tweets, words or segments, while segments have been proven to be advantageous over both words and tweets in news event detection. We study segment-based news event detection, for which existing heuristic-based methods suffer from low recall. We propose feature-based event filtering to address this issue. Our filter incorporate a rich family of features that are empirically proven to be valuable. Experimental results show that our event detection system outperforms the state-of-theart baseline with doubled recall and increased precision.