Assessing Viewer's Mental Health by Detecting Depression in YouTube Videos

TL;DR

This paper presents a machine learning approach to detect depression in YouTube videos by analyzing transcripts and comments, achieving 83% accuracy and supporting mental health assessment.

Contribution

It introduces a novel method combining transcript analysis and comment-based validation to identify depressive content in videos.

Findings

Detection accuracy of 83% for depressive videos

Validation using CES-D scores from comments

Alignment with UN SDG 3.4 for health and well-being

Abstract

Depression is one of the most prevalent mental health issues around the world, proving to be one of the leading causes of suicide and placing large economic burdens on families and society. In this paper, we develop and test the efficacy of machine learning techniques applied to the content of YouTube videos captured through their transcripts and determine if the videos are depressive or have a depressing trigger. Our model can detect depressive videos with an accuracy of 83%. We also introduce a real-life evaluation technique to validate our classification based on the comments posted on a video by calculating the CES-D scores of the comments. This work conforms greatly with the UN Sustainable Goal of ensuring Good Health and Well Being with major conformity with section UN SDG 3.4.