MovieMatcher: A Heuristic to Match Movies Based on Metadata from Different Data Sources

Abstract

With the advent of Web 2.0 and the behavior change which it brought, there are millions of users worldwide contributing to different databases with various forms of data, such as movie ratings, for example. Moreover, the same real-world object (a song, a band or a movie) can be modeled using different ontologies or represented in different ways within the same ontology. Thus, the same film is often described by different attributes in different databases, making it difficult to perform an automatic mapping between those databases. We propose MovieMatcher, which is a heuristic that matches films across different databases using their metadata. After performing 2 experiments with the attempt to match 500 films to IMDb and Rotten Tomatoes databases, MovieMatcher had a success rate of 97.4% and 94.1%, in contrast to an alternative, simpler approach (title exact matching), which had a success rate of 80.8% and 81.9%, respectively.