Multimodal Sentiment Analysis of Social Media Data for Predicting Box Office Outcome

The movie industry is growing at an exponential rate and its success ratio is of utmost importance. It would be immensely beneficial if the outcome of a movie can be predicted well in advance. This would also have an impact on the movie marketing strategies owing to the increasing reliance on the social media. Early prediction of a movie’s outcome is not only a challenging but also relevant in today’s era of Artificial Intelligence (AI) and Machine Learning (ML). This paper makes use of Sentiment Analysis (SA) to examine the performance of movies, using Multimodal data obtained from You Tube. We have compared the performance of three (03) ML classifiers i.e., Random Forest (RF), Support Vector Machine (SVM) and Ada Boost to predict the outcome of a movie by curating a data set having sixteen hundred (1600) plus movies and presented the experimental results and observations. The comparative analysis depicts that RF is the most suitable method for prediction of the Box Office (BO) outcome.