Regression Analysis & Visualization of Twitch Dataset

Abstract

This paper examines Twitch dataset in order to determine and unveil various aspects and relationship among attributes of dataset. Twitch is an online platform used for video game live streaming. The models used by authors in their analysis are Logistic regression, Naïve Bayes, K-Nearest Neighbour and Decision tree. As demonstrated by the authors, K-Nearest Neighbour and logistic regressions outperform other model by yielding an accuracy of 97.7% and 97.4 %.