Prediction of Depression and Anxiety based on Machine Learning

Abstract

Depression and anxiety are prevalent and often comorbid mental health disorders, necessitating accurate prediction and timely intervention to mitigate their impact. This study focuses on leveraging advanced machine learning techniques to predict the levels of depression and anxiety, while also exploring the intricate correlation between these two conditions. Using a dataset comprising depression and anxiety scores, several Machine Learning models, including the SVM, KNN, XGBoost and CatBoost were employed to develop predictive models. The experimental results revealed that CatBoost outperformed other models, achieving an accuracy of 98.9%, followed closely by SVM. Additionally, Pearson and Spearman correlation analyses demonstrated a strong relationship between depression and anxiety scores, with coefficients of 0.6261 and 0.6208, respectively. These findings underscore machine learning’s capacity to forecast mental health issues and highlight the significant correlation between depression and anxiety, providing a robust basis for developing more effective and timely early intervention strategies to improve mental health outcomes.