Abstract by Megan Louder
Predicting Box Office Revenue
The purpose of this research is to predict weekly domestic box office revenue of movies by modeling and predicting average revenue per theater and the number of theaters showing the movie. With these predictions we hope to be able to advise movie theaters and movie studios on the potential of a movie earning more revenue if left in theaters longer. Using data taken from Box Office Mojo, we created a generalized basis function regression model to be applied to each movie in the dataset. Simulation studies were performed to explore the impact of the choice of knot amount and location. The preliminary results of this model seem to predict more accurately and with smaller intervals than a previous ARIMA model chosen. Our research will continue to explore this model and potentially use more covariates such as genre or MPAA rating, as well as compare it with other model choices.