Anxiety, Depression and Stress prediction among College Students using Machine Learning Algorithms

Anxiety, Stress and Depression have become common psychological health issues in today's life. While these health issues have severely affected every age group of people, large number of students are suffering from these issues. The most surprising thing about these issues is that people suffering from them cannot figure out if they have one of these health problems. In this paper, we predicted anxiety, stress and depression using machine learning algorithms. In this paper, we predicted the severity levels of anxiety, stress and depression in college students using machine learning algorithms. DASS21 was used to collect data from 400 students. It is a standard questionnaires used to measure the common signs of anxiety, stress and depression. The severity levels were mild, normal, moderate, severe and extremely severe. The classification algorithms that were applied are Support Vector Machine, KNN, logistic regression, decision tree and naive Bayes. Different calculation matrices like accuracy, precision, specificity were used to compare the models. F1 score measure was included because it was found that the classes were imbalanced in the confusion matrix. Thus it helped find the best model for prediction of these psychological problems. After implementing all the algorithms, we found that K-Nearest Neighbour performed best followed by logistic regression.