Implementation of Explainable AI in Mental Health Informatics: Suicide Data of the United Kingdom

Attempted suicide is a major public health problem that kills a substantial number of individuals worldwide each year. Annually, approximately 800,000 individuals attempt suicide throughout the world. Suicide prevention is usually assumed to include the assessment of risk and preventive variables, as well as associated actions. There are three types of preventive interventions: “universal” (reaching all populations), “selection” (identifying higher-risk groups), and “recommended” (protecting individuals). In this research work, the main aim is to find the predictors which have more contributions to increasing the suicide rate in a specific area. For this purpose, we have analyzed suicide prevention data and have leveraged different machine learning models to predict suicide rates with XAI model SHAP to explain the model. We have also observed the waterfall plot to explain the individual predictors.