Supervised machine learning in predicting depression, anxiety and stress using web-based big data: preserving the humanistic intellect

• Main Authors: Mohammad Aidid, Edre, Musa, Ramli
• Format: Conference or Workshop Item
• Language: English; English
• Published: 2021
• Subjects: RA644.3 Chronic and Noninfectious Diseases and Public Health; RC0321 Neuroscience. Biological psychiatry. Neuropsychiatry
• Online Access: http://irep.iium.edu.my/90102/1/ID%20166%20POSTER%203WCII.pdf; http://irep.iium.edu.my/90102/7/Abstract%20Book%203WCII.pdf; http://irep.iium.edu.my/90102/
• Institution: Universiti Islam Antarabangsa Malaysia

Introduction; The validated Depression, Anxiety and Stress Scale, 21 items (DASS-21) offers an insight on categorizing individuals into severity of each condition. The advancement in public health big data provides a platform for early detection and prompt treatment of individuals. However, there are lacking evidence on prediction accuracy of these data using artificial intelligence methods. Objectives 1. To determine accuracy of supervised machine learning in predicting depression, anxiety and stress using big data. 2. To determine the most important predictor of depression, anxiety and stress using machine learning model Method; Cross sectional study using secondary data obtained from an online resource center was conducted, involving 339,781 respondents. Outcomes were depression, anxiety and stress were measured using DASS21. Each outcome was modelled with the rest of the outcome, plus gender, age, race, marital status, education level and occupational status. Feed-forward artificial neural network was modelled using multilayer perceptron machine learning procedure using IBM SPSS version 2