Leveraging NLP, ML, and Social Media Analysis to Address Mental Health Challenges Among College Students

Authors: Anika Bokka, Renu Balyan

SUNY Campus: SUNY Old Westbury

Presentation Type: Poster

Location: Old Union Hall

Presentation #: 51

Timeslot: Session C 1:45-2:45 PM

Abstract: Healthy Minds National reported that for 2022-23 school year, 36% of college students in the United States were diagnosed with anxiety and 30% with depression/other mood disorders and 65%-87% faced barriers receiving mental health support. The objective of this study is to utilize natural language processing (NLP) to analyze social media data (e.g., Reddit, Instagram, and X), to identify and understand the factors and root causes that contribute to the mental health issues such as anxiety and depression prevalent among university students. We propose to utilize freely available datasets (e.g., Dreaddit) and collect new data by scraping and analyzing social media posts and metadata to find features specific to university students that are correlated with symptoms of psychological disorders (depression and anxiety). This data will be used to develop and train machine learning (ML) models to identify factors impacting mental health and predict probable cases to suffer from these issues in the future. We will employ ML algorithms, including supervised learning methods such as logistic regression, support vector machines (SVM) and others, unsupervised learning techniques including clustering and anomaly detection. Our approach aims to provide a proactive method for colleges and universities to support mental health by identifying students who may be at risk and offering timely interventions. In taking advantage of NLP and ML, institutions can better understand the underlying causes of mental health in their community and implement effective strategies to address them, ultimately fostering a healthier