Speaker: Nguyen Tho Khiem & Fabiana Sabatini , LUISS
Title: The prediction for a popular movie using Bayesian analysis
Abstract: The paper uses Bayesian linear regression method to predict attributes that make a movie popular. The data for the analysis is taken from the Rotten Tomatoes and IMDb websites with more than 600 movies released before 2016. Due to the large number of possible model combination with 16 predictors, the Markov Chain Monte Carlo (MCMC) method is employed with Zellner-Siow Cauchy distribution as prior probabilities for regression coefficients. With BAS package implemented in R, the Bayesian Model Averaging used in the paper shows that Rating on IMDB and Critics score on Rotten Tomatoes websites are 2 most significant characteristics predicting the popularity of a movie. Finally, we discuss the limitation of the employed method in the paper and the direction for further analysis.