IMDB Spoiler Dataset
Abstract
User-generated reviews are often our first point of contact when we consider watching a movie or a TV show. However, beyond telling us the qualitative aspects of the media we want to consume, reviews may inevitably contain undesired revelatory information (i.e. 'spoilers') such as the surprising fate of a character in a movie, or the identity of a murderer in a crime-suspense movie, etc. In this paper, we present a high-quality movie-review based spoiler dataset to tackle the problem of spoiler detection and describe various research questions it can answer.
Turn this paper into a lesson
ArcXiv compiles a structured reading guide from this paper's metadata: plain-English importance, contributions, prerequisite concepts, which sections to read first, flashcards, and a quiz. Grounded in the abstract, never invented.