Computers Beat Film Critics

Computers are better at figuring out which movies will stand the test of time and which will die out, according to a study by Northwestern University.

The best predictor of a movie's significance, as per the study, is how often it is cited by other movies. Computers, the research found, are better at identifying a movie's greatness than film critics, who have bias and may be overconfident about certain types of films at times.

The method by which the study evaluated films is entirely objective, in contrast to the film critic's method. The researchers considered subjective metrics such as critical reviews, awards and public opinion as well as objective metrics like box office sales and citations.

Particularly for movies that are 25 years or older, the researchers found that their software program was better at predicting greatness and beat film critics, the wisdown of the crowd, number of awards won and box office sales.

The study, published in the Proceedings of the National Academy of Sciences, is an analysis of 15,425 films produced in the United States and listed in the Internet Movie Database (IMDb). The researchers tried to determine whether a movie would appear in the National Film Registry of the U.S. Library of Congress. The number of times a movie is referenced by other movies appeared to be the best predictor of whether it would be included in the Registry.

Some films with the most citations that are also in the Registry are The Wizard of Oz, Star Wars, Psycho, Casablanca and Gone With the Wind.

The automated method can be applied to older films for which other metrics may not be available. Luís Amaral, the leader of the study and co-director of the Northwestern Institute on Complex Systems, plans to use this objective analysis method to evaluate the importance of scietific research papers, paintings and music.

More than 1 million scientific papers are published each year.

To contact the writer, email: sonali.raj@gmail.com