Design and implementation of computational models for film assessment and classification

Resumen

Machine Learning is impregnating every industry and every area of our lives, so it made
sense that it was applied to a field as linked to technology as cinema. Big production
companies as Netflix have already started to use algorithms to find ways of telling stories
that are more appealing to the audiences.
But good films still fail in the box office, so this project tackles the supposition of
assessing the quality of a film, based on the standards of a particular film festival, in
order to try to predict the failure or success a film will have. This project will work in
a restrained environment but the suppositions made could be extrapolated to feature
films, with the proper resources and time.
We will analyze what features make each film unique, what descriptors can give us
hidden information about the video and the audio and how this features can be fed to
computational models to predict weather a film will end as Finalist or not of a given
film festival.