Predicting Success of Bollywood Movies Using Machine Learning Techniques

An enormous amount of Bollywood movies are released every year making Bollywood one of the largest film industry in the world. In this study, we apply machine learning tools to create a model which can predict whether a Bollywood movie will be successful or not, before it is released. We have collected data from multiple sources like Cinemalytics, BoxOfficeIndia, YouTube and Wogma. We have designed factors like music score which is a unique element to Bollywood movies and greatly increases the accuracy of prediction. We label the prediction in two classes, Hit and Flop. After evaluating multiple classifiers, we have used Bagging algorithm to create the model.