Machine Learning Applications in Early Screening of Depression and Anxiety using RCADS-47

Abstract

Depression and anxiety are prevalent among 10-20% of children and adolescents globally, with an estimated 15 million people affected in Pakistan. Despite this growing figure, the general Pakistani population lacks awareness regarding mental disorders due to limited mental healthcare resources and negative perception of mental health. This study aims to utilize machine learning with RCADS to maximize the use of current healthcare resources and facilitate depression and anxiety screening. Three feature selection methods i.e., the Chi-square test of independence, Spearman correlation, and Recursive Feature Elimination revealed a weak correlation with the evaluation of depression and anxiety in the study population. Data augmentation was done using the multinomial probability distribution of the existing data to generate hybrid-synthetic correlated discrete multinomial variates of each item of RCADS-47, to address the limitation of a small sample size. Six commonly employed ML algorithms—Decision Tree, Random Forest, Support Vector Machine, Logistic Regression, Naive Bayes, and K-Nearest Neighbor—were trained on the hybrid data to develop the predictive models. The Naive Bayes algorithm yielded the best overall results with up to 75% accuracy and a 0.75 F1 score. The findings suggest that the Naive Bayes algorithm using 46 features suits the data well and has the potential to be used as a data-driven decision support system for the concerned professionals and improve the usual way of screening anxiety and depression in children and adolescents.