Abstract
The paper discusses the existence of distinct communities of fine-grained semantic features in movies, which may result in some movies being popular at box office, others winning Oscars while some receiving high ratings from users/critics. We study a phenomenon related to the Movie Genome Project, which aims to categorize movies by user taste, mood, story, plot development and other semantic meta-data. More specifically, our research reveals that four unique communities of semantic genomes appear in every successful movie. Similarly, five unique gene communities describe very poorly rated movies. Using this community structure of genomes representing the network of semantic features in movies, we develop an optimization function that attempts to identify the genetic algorithm powering successful or profitable movies. Our results indicate that utilizing movie genome communities in genetic optimization perform better than standard classifiers such as decision trees in predicting movie profitability.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
