SUICIDAL BEHAVIOR PREDICTION USING DATA MINING TECHNIQUES

Background: Suicide is one of the most serious public health problem that has affected many people. After being recognized as a public health priority by the WHO (World Health Organization) various studies have been going out for its prevention. It is one of a serious health problem and it is preventable and can be controlled by proper interventions and study in the field. The objective of the study is to create a prediction model for individuals who are at higher risk of suicide by studying the different predictors of suicide such as depression, anxiety, hopelessness, stress etc. by using data mining techniques for the prediction. Study Design: Systematic review and predictive analysis for suicidal behavior. Methods: The research applies data mining process to analyze the data and on the basis of analysis create the model to predict suicidal behaviors present in the individual. Prediction is done on the basis of analysis of risk factors which are Depression, anxiety, hopelessness, stress, or substance misuse which is calculated by using various psychological measures such as Beck hopelessness scale,suicidal ideation subscale,hospital anxiety and depression scale.Various data mining algorithms for classification are compared for the purpose of prediction. Results: Six different data mining classification algorithms which are namely Classification Via Regression, Logistic Regression. Random Forest, Decision Table, SMO are compared and Classification Via Regression was found to the highest accuracy in prediction. Conclusions: Data required for the development of such a model requires continuous monitoring and needs to be updated on a periodic basis to increase the accuracy of prediction.