Effects of film characteristics on consumer behavior; A text classification approach

Abstract

Film industry has lots of consumers worldwide. Studying these consumers behavior is one of the interesting subjects in business management and other disciplines like computer sciences. This paper analyzes the effects of film characteristics on audiences by text classification based algorithms. These algorithms are designed to classify the documents into classes automatically. The main data base is extracted from IMDB, which is a popular online film information database. By extracting the film features as texts and giving them as an input to classification algorithms such as SVM algorithm, we show the relationship between these features and consumers behavior which is shown by audience’s ratings. Features such as director, stars and genre are most focusing features that are given to the algorithms. By focusing on Oscar winner films we extracted some other detailed features and find their trace on audience’s behavior. We also compare the features effect with each other and find the most important groups of them.