A Binarization Approach for Predicting Box-Office Success

Authors

  • S Jahangeer Sidiq, Majid Zaman

Abstract

Film industry is the largest industry where thousands of movies are released every year. Predicting box-office success of a movie prior to its release is an interesting topic as large amount of capital is involved .Predicting success of movies using historical data, social media data using machine learning algorithms is very beneficial for all the stake-holders in film industry. In this paper we use a data set imported from UCI machine learning repository, machine learning classifiers (KNN, Naïve Bayes, Decision Tree) and some binarization algorithms (OVO, OVR and ECOC). We also used Bagging technique for model creation that shows the best results independent of which base learner is used.

Keywords-Box-office; Machine learning; Classifiers; Ensembles; Binarization.

Published

2020-12-31

Issue

Section

Articles