Suicide Prediction in Twitter Data using Mining Techniques: A Survey

Online social networks and the World Wide Web specifically Twitter, have expanded the network between individuals with an end goal that data can be spread to a huge number of individuals in a matter of minutes. This type of online collective contagion has given numerous advantages to society, for example, giving consolation and crisis administration in the quick fallout of catastrophic events. Be that as it may likewise represents a potential hazard to powerless Web clients who get this data and could consequently come to hurt. One example is the spread of suicidal ideation in online Social network like Twitter, which has risen in concern. In this Survey paper, we report the consequences and results of various machine classifiers worked with the point of arranging text identifying with suicide on Twitter. The classifier recognizes the more stressing content, for example, suicidal ideation, and other suicide related points, for example, detailing of a suicide, dedication, campaigning and bolster. It aims to refer and identify flippant references to suicide by using baseline classifier using emotive, lexical, psychological, and structural features from twitter. In this survey clustering, classification, martingale function, association rule, Natural language processing (NLP) and different machine learning techniques are used. This paper further discusses about the research challenges in this domain that initiates the scope for further research.