Movie clustering using the self-organizing map: MOVSOM 2004-06-19

Abstract

The aim of this report is to investigate whether it is possible to determine whether two or more movies are alike. We will describe two approaches; one where we only use the movie's plot description and one where we use preselected keywords that are significant for the movie. Each movie is represented as a vector in the input space. If two or more movies are similar then their (Euclidian) distance in input space will be small which will result in clusters on a 2D-SOM grid. By studying the 2D-SOM grid we hope to be able to determine whether two or more movies are alike.

Bibtex

@unpublished{gabrielsson04movsom-i,
  author={Stefan Gabrielsson and Sam Gabrielsson},
  title={{M}ovie clustering using the self-organizing map: {MOVSOM}},
  note={Uppsala University, Department of Information Technology},
  year={2004}
}