News Popularity Prediction with Machine Learning

Online news is quite popular as it offers valuable information quickly at no cost. However, it can be challenging to write news articles that can get immediate attention from the viewers. Journalists and writers adopt different approaches to writing similar content. However, only a handful of stories receive much attention and are further shared. The probability of an article being circulated widely online depends on various factors. Hence in this work, we predict the popularity of news articles. We formulate the problem as a regression problem and use nine different regression techniques to provide an optimal recommender system. Additionally, we assess the effectiveness of different feature reduction techniques and apply robust feature scaling. Our analysis shows that Gradient Boosting outperforms the other techniques.