A First Dataset for Film Age Appropriateness Investigation

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 1311–1317 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 1311 A First Dataset for Film Age Appropriateness Investigation Emad Mohamed, Le An Ha RGCL, RIILP, University of Wolverhampton Wulfruna Street, Wolverhampton, WV1 1LY, UK E.Mohamed2@wlv.ac.uk, L.A.Ha@wlv.ac.uk Abstract Film age appropriateness classification is an important problem with a significant societal impact that has so far been out of the interest of Natural Language Processing and Machine Learning researchers. To this end, we have collected a corpus of 17000 film transcripts along with their age ratings. We use the textual contents in an experiment to predict the correct age classification for the United States (G, PG, PG-13, R and NC-17) and the United Kingdom (U, PG, 12A, 15, 18 and R18). Our experiments indicate that gradient boosting machines beat FastText and various Deep Learning architectures. We reach an overall accuracy of 79.3% for the US ratings compared to a projected super human accuracy of 84%. For the UK ratings, we reach an overall accuracy of 65.3% (UK) compared to a projected super human accuracy of 80.0%. Keywords: Machine Learning, Deep Learning, XGBoost, Age Appropriateness, Movies 1. Introduction Interest in film from a computational linguistics per- spective has been massive. Several studies in CL have examined the genre in terms of Sentiment Analysis (Phan and Matsumoto, 2018), turn-taking (Banchs, 2012), and many other things. Most of these studies have focused on film reviews (from Amazon.com and IMDB, but the actual film content (the script, audio, and video for example) has not received as much interest in spite of the potential availability of huge datasets as will be demonstrated below. In this study, we investigate a completely new problem with a novel corpus. We investigate whether we can use Machine Learning and Artificial Neural Networks to predict the age classification accompanying the films based on their textual content. For this purpose, we collect a custom corpus from the internet and complement it with age rating certificates from the Internet Movie Database (IMDB). The data combination makes a dataset that can be used for many purposes part of which is the focus of this paper: age appropriateness classification. Entities like the Motion Picture Association of America (MPAA)1 and the British Board of Film Classification (BBFC)2 issue film ratings that determine the age appro- priateness of each film based on the film’s content. The lat- ter define their classification as ‘the process of giving age ratings and content advice to films and other audiovisual content to help children and families choose what’s right for them and avoid what’s not.’ This is a human labour intensive endeavor that requires at least two Compliance Officers to watch each film and report on it. So far, there does not seem to be enough interest from the Computa- tional Linguistics community in the problem of automatic age appropriateness classification, possibly due to the lack of resources. We have collected a fairly large …

Content truncated to 500 words