Advanced machine learning models for Depression level categorization using DSM 5 and personality traits

Abstract

The COVID-19 pandemic has led to an increase in mental health problems, such as depression. Depression is a major cause of disability and can affect anyone. It is important to seek early detection and treatment for depression, as it can have a significant impact on a person’s life. This research paper explores the complex and multifaceted nature of human personality using the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) criteria and Eysenck’s personality traits questionnaire. The paper then outlines the methodology employed, involving the collection of data through surveys consisting of 66 questions divided into six forms. DSM-5 criteria are applied to assess depressive episodes, while Eysenck’s questionnaire evaluates personality traits. The paper’s approach involves two stages: predicting personality traits and estimating stress levels. Multiple machine learning models are utilized for these predictions. The results of this study highlight the effectiveness of the Support Vector Machine (SVM) classifier, which consistently outperforms other models, achieving impressive accuracy in predicting both personality traits, such as extroversion and neuroticism, and stress levels. Notably, SVM demonstrates its prowess with an accuracy of 91.43% in predicting stress levels.