Detecting Causality by Data Augmentation via Part-of-Speech tagging

Abstract

With a deluge of text-based data available, the ability to automatically extract important information from the text data is crucial, especially extracting events from factual text data like news articles. Finding causal relations in texts has been a challenge since it requires methods ranging from defining event ontologies to developing proper algorithmic approaches. In this paper, I developed a framework which classifies whether a given sentence contains a causal event. As my approach, I exploited an external corpus that has causal labels to overcome the small size of the original corpus (Causal News Corpus) provided by task organizers. Further, I employed a data augmentation technique utilizing PartOf-Speech (POS) based on my observation that some parts of speech are more (or less) relevant to causality. My approach especially improved the recall of detecting causal events in sentences.