Predicting Factors Influencing the Suicides in Sri Lanka

Abstract

Suicide is a long-term social issue and a common cause of unnatural death. An individual's suicide risk is usually determined by mental health, but it is also influenced by their background. This research focuses on identifying the factors influencing suicide by scanning the civil, educational, and professional backgrounds of Sri Lankans who have committed suicide from 2014 to 2019. The factors considered in the study are Age Group, Gender, Civil Status, Education Level, Nature of the Occupation, and Reason for Suicide. Initially, the data set is clustered using the k-mode algorithm and identified five clusters that are centered on five different reasons of suicide. Next, the Apriori algorithm is used to identify the associations between the attributes which could lead someone to suicide due to a particular reason. The algorithm is applied for both the entire data set and each cluster. The rules mainly generated around five reasons such as mental disorders, addiction to narcotic drugs, chronic diseases, problems caused with the elders, and harassment by the husband and family disputes. The final evaluation showed that the identified rules are correct with 74%.