Prediction of Suicide Attempts Using Machine Learning

Abstract

In Bangladesh, there is a rising public health issue around suicide. Targeted suicide control and prevention efforts can be greatly aided by a thorough understanding and forecast of suicide patterns. In this research, after initially analyzing suicide trends and geographical distribution of suicides in Bangladesh, we developed a machine learning model for predicting suicide at the county level for the seven-year period between 2015 and 2022 using publicly available data. Suicide is the intentional act of harming oneself in an effort to end one's life. Suicides frequently have a variety of causes, including despair, financial hardship, mental illness, legal issues, encompassing situations, etc. People's propensity for self-destruction may be a severe problem that is not unique to any one state or nation. Suicide may significantly affect the worldwide death rate, according to data at the global level. Additionally, suicide is among the top twenty causes of death in the globe. Suicide is a subject that has gained increasing cultural attention. Actually, it is one of the major causes of death in the contemporary world. In order to counter this threat, it is crucial to develop precise prediction algorithms based on available data. The research primarily examines suicide data, finds significant risk variables, and forecasts future suicide attempts with a high degree of accuracy. Three machine learning algorithms—logistic regression, random forest, and Naive-Bayes—have been compared with the goal of predicting suicide.