Can we predict the Oscar winner? A machine learning approach with social network services

Abstract

The Academy Awards are one of the most prestigious events in the global movie industry, and thus, the Oscar winner has been one of the most remarkable and hottest topics in society. Due to this, many media companies attempt to predict which nominated film will win the Oscar award (Academy Award for Best Picture). Moreover, public perceptions of each film are among the most important indicators of the Oscar award. In light of this, the present study examines user-created posts about nominated films on a topic-oriented social network service. After sentimental differences in the posts about the Oscar winners and other nominated films are examined, the concept of the adjusted Oscar winner score is proposed, explored, and tested. The results indicate notable sentimental differences, while the score is well computed and effective for computing the 2020 Oscar winner. Based on the results, the implications and key limitations are discussed.