Machine Learning-based Classification of Persistent Depressive Disorder using Questionnaire Responses

Detecting stress is crucial for monitoring mental health, and recent advancements in machine learning have enabled the automation of this process. This paper proposes a novel approach for stress detection by combining unstructured Twitter data with structured questionnaire responses. The structured data comprises answers to stress-related questions covering various behavioral and emotional aspects, while the unstructured data consists of tweets potentially expressing stress-related sentiments. The methodology involves preprocessing the data, extracting features, and training a machine learning model, such as a Support Vector Machine, using the structured questionnaire responses. This trained model is then applied to predict stress in the unstructured Twitter data. Performance metrics such as accuracy, precision, and recall are used to assess the model's effectiveness. The development of this stress detection model offers a valuable tool for automated mental health monitoring, providing insights into stress prevalence within specific groups. Furthermore, the study demonstrates the potential of integrating structured and unstructured data sources to enhance predictive modeling. Given the significant impact of stress on contemporary culture, effective methods for its identification could lead to improved health outcomes and early intervention. Additionally, the article includes a section on stress management strategies to broaden the project's scope.