Event Extraction from social media Text in Malayalam using Neural Conditional Random Fields

This paper describes a Neural Conditional Random Fields (NCRF) approach for Event extraction (EE) task which aims to discover different types of events along with the event arguments from the user generated text content (tweets) in Malayalam. The data for this work was obtained from FIRE (Forum for Information Retrieval and Evaluation) 2017 shared task [12] on Event Extraction from Newswires and Social Media Text in Indian Languages. A NCRF is a combination of Recurrent Neural Network (RNN) and Conditional Random Fields (CRF). In addition to event detection, the system also extracts the event arguments which contain the information related to the events such as when (Time), where (Place), Reason, Casualty, Aftereffect etc). Our proposed Event Extraction system achieves F-score of 79.74%. The results are encouraging and comparable with the state-of-art.