Comparison of Machine Learning Models to Predict Twitter Buzz

Abstract

This paper investigates six machine-learning models to determine which algorithm would effectively predict buzz on Twitter. Different classifiers are applied such as Stochastic Gradient Descent, Support Vector Machines, Logistic Regression, Deep Neural Networks, Random Forests and Extra Trees on a Twitter dataset. This dataset contains features with users and author engagement over a certain period. After tests conducted on all the algorithms, we concluded that Extra Trees model outperforms the other models.