Semantic-based Social Media Threats Detection

Abstract

The exponential growth of internet access and computing devices, has led to the growth of social media usage and changed the way of our lives in society. Due to the growth of social media usage such as Facebook, Twitter, Instagram, YouTube, etc., there are possibilities that users share information which may pose a threat to the national security. For example, the threats might be propagated by a certain group of terrorists. The threats can be written either in Malay or mixed with other languages; especially English and in a short form basis. The major issue is to extract the 'meaning' of the messages for further processing tasks (e.g. categorizing and identifying the insight of threats). Therefore, a semantic-based grammar approach helps to fill this 'knowledge gap' for threat analysis at the early stage. This research will focus on knowledge extraction (based-on Malay words) through semantic relations between words and their dependents.