Aseneth Jepchirchir, Obadiah Musau, Mvurya Mgala, Fullgence Mwakondo.

ABSTRACT

The proliferation of social networks and smartphones has fundamentally changed how we consume and engage with online content. Previous studies has shown that the increased exposure to suicide stories through social media is associated with the increase in suicidal ideation across all genders. In this study, we present a systematic review of literature on previous machine learning research in suicide ideation and prediction. The research will use the Snowballing methodology to review the literature. This will involve identification of significant and recent literature on predictors of suicidal behaviors using keywords such as adolescence, social media, prediction, data sets, and mental health. Tracking of references and citations will be conducted through backward and forward snowballing. To ensure efficiency and relevance of the reviewed papers to the topic of study, application of Boolean operators will be utilized to reduce the search space. The review will make use of online databases including; TUM University catalogue, Project MUSE, Medline, and Google scholar. The finding of this study can be used to design a prediction model for suicidal ideation and behaviors to be used by primary health workers for early detection and intervention.

Key words: Predictor, Suicidal thoughts, Adolescent girl, Social media