Leveraging Machine Learning to Analyze Social Media Use and Understanding the General Behavior in ADHD Communities

Abstract

Social media is an online platform that allows people to connect, share information, and create communities. It is a way for individuals, groups, and organizations to interact and engage with one another through the internet. ADHD (attention deficit hyperactivity disorder) also known as one of the most common neurodevelopmental disorders of childhood, mostly diagnosed in childhood and lasts into adulthood. People with ADHD have trouble paying attention, controlling their impulsive behavior, and being overly active. ADHD and social media often associate with one another because social media users are more likely to experience new ADHD symptoms. The symptoms are frequently seen in adults who often use social media and they tend to have short attention spans. This study interpreted a survey of active social media users on a likert scale involving the hours of social media usage and questions directed to ADHD symptoms. Classification method is used to elucidate how social media can be used to see how likely someone to have ADHD based on the survey. The result is the link between social media use and ADHD in individuals and which classifier would be the best to be used. Using Classification type of models as we are using supervised learning, we conclude that Support Vector Machines (SVM), Logistic Regression (LR), and Gaussian Naïve Bayes have the stable performance without overfitting or underfitting conditions compared to Random Forest (RF) that have the highest accuracy in training but then dropped on validation and testing leading to potential overfitting conditions.

Trailer Video: https://youtu.be/cPz5yscS4gU?si=ULGqkwUhgaJmUWeN

Presentation Video: https://youtu.be/oO_nsHJvb5Q?si=_MkCtHW-tRLe6Ug6