Assessment of Anxiety, Depression and Stress using Machine Learning Models

Over the last few decades, psychological health issues have become very common in people worldwide. In this paper, prediction of the occurrence of psychological problems such as anxiety, depression and stress has been made by applying eight machine learning algorithms to data taken from the online DASS42 tool. Five different severity levels of anxiety, depression and stress have been predicted using eight algorithms. The algorithms are grouped into four categories: probabilistic, nearest neighbor, neural network and tree based. A hybrid classification algorithm was also applied for prediction of different severity level anxiety, depression and stress. The same methods were also applied to another dataset, DASS21 collected by authors. The prediction accuracy found by using the hybrid algorithm was greater than by using single algorithms, but the highest accuracy was found by use of the radial basis function network, which comes under the category of neural network.