Suicidal Ideation Detection: Application of Machine Learning Techniques on Twitter Data

Abstract

The World Wide Web, particularly Twitter, and online social networks have expanded the network connecting people, allowing for the rapid dissemination of information to large numbers of people. There are several instances of this kind of online collaborative contagion, one of which is the development of self-destructive ideas in social media sites like Twitter, which has caused alarm. In this investigation, the implications and findings of several machine classifiers that were applied to the point order of tweets and terms connected to suicide are discussed. The classifier can distinguish between more stressful information, such as suicidal creativity, other suicide-related topics, in-depth suicide-related facts, loyalty, campaign, and support. A simple classifier utilizing emotional, lexical, psychological, and structural characteristics from Twitter is used to link and identify allusions to suicide. This procedure makes use of clustering, bracketing, association rules, NLP (natural language processing), and numerous machine-learning techniques. This research study explores the restrictions or difficulties in this field and serve as a guide for future research.