Data Mining Approach to Understand the Association Between Mental Disorders and Unemployment

摘要

Over the years, mental illness has affected the life of numerous human beings and nowadays is a matter of great concern. The problems that arise with this clinical condition, such as social isolation, unemployment, and others, have been a subject of study. The purpose of this study is to use a survey that aims to assess the situation of unemployment among individuals with mental illness. Hence, this article focuses on using the result of this research to identify if there is a connection between having mental illness and being in a situation of unemployment, as well as, which factors can be determinant for such a relationship and also if there is any way to anticipate them. In this context, this research attempts to develop an accurate prediction mechanism, using Data Mining, capable of predicting, based on the answers of a similar questionnaire, if an individual will be in risk of unemployment. Throughout this research, the CRISP-DM methodology was adopted and the RapidMiner Studio software was the tool used for the learning process. The best percentages of accuracy were between 0.79 and 0.86, of sensitivity between 0.75 and 0.88, and of specificity between 0.66 and 0.93.