Attribute Selection and Classification on Mental Illness DataSet

Abstract

Mental illness is increasing rapidly through the world where people are now susceptible to mental disorders at some point in their lives. Mental disorders can also be a major cause of global disease burden over the coming year since the awareness among the world community on the importance of preventing mental disorders is still low. Thus, initial awareness of mental disturbances should be enhanced by identifying the early symptoms of the disease. By having early identification, people will tend to become more sensitive to prevent mental disorder. In this research, Mental Illness dataset that used are downloaded from Kaggle website that measures attitudes towards mental health and frequency of mental health disorders in the technology workplace. Mental illness dataset will be tested by using attribute selection techniques which is GainratioAttributeEval, CorrelationAttributeEval, CfsSubsetEval, InfoGainsubsetEval, and Wrapper selection in WEKA tool to identify the important attribute by removing the redundant and irrelevant attributes.  The performances accuracy of the attribute selection was validated by using several classification techniques. There are four classifiers used to evaluate the performances such as Naïve Bayes, K-Neighbours Nearest, Decision Tree and Logistic regression.