Popularity prediction of movies: from statistical modeling to machine learning techniques

Film industries all over the world are producing several hundred movies rapidly and grabbing the attraction of people of all ages. Every movie producer is of keen interest in knowing which movies are either likely to hit or flop in the box office. So, the early prediction of the popularity of a movie is of the utmost importance to the film industry. In this study, we examine factors inside the hidden patterns which become movie popular. In past studies, machine learning techniques were implemented on blog articles, social networking, and social media to predict the success of a movie. Their works focused on which algorithms are better at predicting the success of a movie but less focused on data and attributes related to an ongoing movie and in various directions. In this paper, we inspect this perspective that might be related to the prediction of the results. Data collected from the publicly available Internet Movie Database (IMDb). We implemented five machine learning algorithms, i.e., Generalized Linear Model (GLM), Deep Learning (DL), Decision Tree (DT), Random Forest (RF), and Gradient Boosted Tree (GBT) using Root Mean Squared Error (RMSE) as a performance metric and got the accuracy performances of GLM: 47.9%, DL: 51.1%, DT: 54.5%, RF: 50.0%, and GBT: 49.5%, respectively. We found that GLM is the high achieving accuracy regression classifier due to the lower value of RMSE, which is considered to be better.