Measuring shared interest in an online movie-rental service

Abstract

Having common interests organizes people in groups. In this work, we introduce a parameter E which measures shared interest of users in movies in an online movie-rental service. The threshold value of E determines the stability of the network, i.e. high values of Eth can form a tightly clustered network robust to targeted removal of movies. Analyzing the time evolution of movie rental behavior shows that the rental frequency of movies is not linear and that when fitted with a power-law, shows a nearly constant scaling parameter of 1.5.