$$\hbox {NE}^2$$ NE 2 : named event extraction engine

Abstract

Named event discovery using news headlines is an important problem with various applications in story telling, news event exploration, social media information fusion, etc. Named events are short phrases that represent the name of events like 2016 Rio Olympic Games, 2G Case, and Adarsh Society Scam. Existing work has largely focused on discovering events of named events using data mining and text mining techniques. However, the problem of discovering named event has not been addressed yet. In this paper, we present a system $$\hbox {NE}^{2}$$ that uses pattern- based method to discover named events using news headlines. Along with named event, we also discover its categories, popular durations, popularity, and type of named events. Named events are categorized into candidate-level and high-level categories using URL information, and popular durations of named events are extracted using temporal information of news headlines. Our system generates 75,689 number of named events by analyzing 6.5 million news headlines. Out of 75,689 named events, 62,950 (82%) are categorized and popular duration are extracted for 73,288 (96.8%) number of named events. Based on performed experiments, our proposed system $$\hbox {NE}^{2}$$ has 68% of accuracy for named events, 71.6% for named event’s category, and 78.4% for named event’s popular duration.