A New Approach for Twitter Event Summarization Based on Sentence Identification and Partial Textual Entailment

Recent trend of information propagation on any real-time event in Twitter makes this platform more and more popular than any other online communication media. This trend creates a necessity to understand real-time events quickly and precisely by summarizing all the relevant tweets. In this paper, we propose a two-phase summarization approach to produce abstract summary of any Twitter event. The approach first extracts key sentences from the whole set of event relevant tweets and eliminates maximum redundant information by exploring Partial Textual Entailment (PTE) relation between sentences. Next, generates an abstract summary over the least redundant key sentences. We conduct experiments to evaluate the performance of our propose approach and report that the approach outperforms over the baseline approach as well as state-of-the-art event summarization approach.