Supporting the Treatment of Mental Diseases using Data Mining

摘要

Mental disorders are a rising phenomenon in Bangladesh. This phenomenon has contributed to intensive psychological healthcare data. It may change into helpful information via data mining application. In Bangladesh, healthcare data is underutilized. There are fifteen million individuals enduring from mental diseases of the many sorts in our country. Particularly, nearly 10 percent of the people seriously required mental health services. Early treatment of mental state issues helps the psychiatrist to treat it as a primary stage. For various mental problem symptoms are similar which makes diagnoses very complex task to recognize and sometimes doctors misjudged the disease. The objective of this research is to examine a classification algorithm to predict mental disorder. In this study, we analyze 466 mental health patients dataset to find the relation between diagnosis and attributes. We applied three machine-learning techniques: Random forest, SVM, K-nearest neighbor and compared performances of the above algorithms using various measures of accuracy to detect mental health problems. Experimental results show that Random forest shows a superior performance than the other algorithms we applied.