Linking Mental Health and Gaming Using C-Support Vectors and Multilayer Perceptrons

Gaming has become an integral part of our lives, transcending ethnosocial barriers and ubiquitous enough for every individual with access to the internet having gamed at some point in their lives. While gaming is commonly considered a recreational activity that reduces stress, this paper maps the link between gaming and stress levels. The paper aims identify how various factors of gaming including but not limited to the game, the type and genre of the game, the age and gender of the individual, and playtime factor into general anxiety disorder levels of the person. Firstly, the paper does some basic data analytics on the dataset using libraries such as pandas and matplotlib to visually identify potential cues. It also proposes two models to find the relationship between these factors and stress levels. The first model proposes is a C-Support Vector Classification model which uses a separating hyperplane to classify the game and the second is a Multi-layer Perceptron classifier which uses LBFGS to classify the dataset.