Age Suitability Rating: Predicting the MPAA Rating Based on Movie Dialogues

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 1327–1335 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 1327 Age Suitability Rating: Predicting the MPAA Rating Based on Movie Dialogues Mahsa Shafaei, Niloofar Safi Samghabadi, Sudipta Kar and Thamar Solorio Department of Computer Science, University of Houston {mshafaei, nsafisamghabadi, skar3, tsolorio}@uh.edu Abstract Movies help us learn and inspire societal change. But they can also contain objectionable content that negatively affects viewers’ behavior, especially children. In this paper, our goal is to predict the suitability of movie content for children and young adults based on scripts. The criterion that we use to measure suitability is the MPAA rating that is specifically designed for this purpose. We create a corpus for movie MPAA ratings and propose an RNN-based architecture with attention that jointly models the genre and the emotions in the script to predict the MPAA rating. We achieve 81% weighted F1-score for the classification model that outperforms the traditional machine learning method by 7%. Keywords: Text Classification, MPAA Rating, Story Analysis 1. Introduction The latest reports on screen time among children show an alarming trend in the amount of time they spend watching movies or videos on electronic devices (Chen and Adler, 2019). Some of the content in these movies and videos is completely innocuous, but there is also harmful and in- appropriate content that can negatively affect their behav- ior. For example, watching specific programs may en- courage irresponsible sexual behavior and alcohol usage in teenagers (Strasburger, 1989; Sargent et al., 2006; AAP, 2001) or instill anxiety and fear among children (Wilson, 2008; Johnson et al., 2002). The Motion Picture Association of America (MPAA)1 is a film rating system that establishes the appropriate age for movie viewers. MPAA ratings have a wide practical value. For example, parents can rely on them as a guideline to determine what movies are appropriate for their children. Also, media service providers (e.g., Amazon and Netflix) may use these ratings to enable age filters in parental con- trols. Having the MPAA rating is an important element for producers too. Although films can be shown without a rating, certain theaters refuse to show non-rated movies2 which in turn negatively affects the potential popularity of the movie as well as its gross revenue. The MPAA rating is determined by CARA3 (one of the subdivisions of MPAA organization). Members of CARA watch the entire film to determine the age category and the MPAA rating of the movie (MPAA, 2010). The current rat- ing method is a time-consuming and non-scalable process. Not surprisingly, there are many movies available that are missing an MPAA rating. Also, rating happens post pro- duction, when making changes in movies can cost a lot of money. To solve the aforementioned obstacles, we explore predict- ing the MPAA rating by only using movie scripts. Our method is a practical solution because it can predict the rat- ing at early steps of the production when we only …

Content truncated to 500 words