Suicidal Risk Evaluation Using a Similarity-Based Classifier

Suicide remains one of the leading causes of death in the world and it is showing an increasing trend. Suicide is preventable by early screening of the risks. But the risk assessment is a complex task due to involvement of multiple predictors, which are highly subjective in nature and varies from one case to another. Moreover, none of the available suicide intent scales (SIS) are found to be sufficient to evaluate the risk patterns in a group of patients. Given this scenario, the present paper applies similarity-based pattern-matching technique for mining suicidal risks in vulnerable groups of patients. At first, medical data of groups of suicidal patients have been collected and modeled according to Pierce's Suicide Intent Scale (PSIS) and then engineered using a JAVA-based pattern-matching tool that performs as an intelligent classifier. Results show that addition of more factors, for example, age and sex of the patients brings more clarity to identify the suicidal risk patterns.