Movie actor key attributes success prediction with network community detection

Though the film entertainment industry has the potential for tremendous worldwide revenues, the prediction of success relies on an enormous number of variables. In order to determine the importance of variables which impact a film's success, a prediction model is needed. One approach is to identify communities within the network to predict a movie's success variables such as revenue, winning awards, and ratings. This study focuses on network clustering to identify communities within a large network of movies and actors. The network investigated in this work is a type of collaboration network in which movies are connected to each other if they share at least one actor together. The results indicate how we can identify and use these communities to determine the key attributes which lead to movies' success. We also demonstrate which genre types are more correlated to communities' topology features (density and size).