A machine learning approach to detect depression and anxiety using supervised learning

Abstract

Depression, a major depressive disorder and anxiety are common medical illness which cause several symptoms that a ect the way a person feels, thinks, and the way he/she acts. These disorders are not only hard to endure, but are also risk factors for heart disease, panic attacks, dementia, and thus causing severe distress and functional impairment.Overall, more than 50% of the general population in middle and high-income countries su ers from at least one of these mental disorder at some point in their lives. In order to detect these disorders at an early age, we have proposed a model that uses a standard psychological assessment and machine learning algorithms to diagnose the di erent levels of such mental disorders. In our proposed model we used ve di erent types of AI algorithms: Convolutional neural network, Support vector machine, Linear discriminant analysis, K Nearest Neighbor Classi er and Linear Regression on the two datasets of anxiety and depression. These algorithms are used to nd the severity level of anxiety and depression, a patient is going through. In this paper we compared the results of the ve algorithms on the two datasets separately on the basis of di erent measurement metrics. The proposed model achieves the highest accuracy of 96% for anxiety and 96.8% for depression using the CNN algorithm while its results were compared with the other ve algorithms we have used for our model.