On the Application of Bayesian Credibility Theory to Achieve More Accurate Movie Rankings
Published: 2016
Author(s) Name: Palash Ranjan Das, Tripti Chakrabarti |
Author(s) Affiliation: Department of Applied Mathematics, University of Calcutta, Kolkata, West Bengal, India
Locked
Subscribed
Available for All
Abstract
Credibility theory is a branch of actuarial science devoted to quantify how unique a particular outcome will be when compared to an outcome deemed as typical. In this paper, we will examine the application of the principles of Bayesian Credibility Theory in rating and ranking movies by a premier online movie database which is based on users votes. Although the Bayesian credibility theory was developed originally as a method to calculate the risk premium by combining the individual risk experience with the class risk experience, it is generic enough to deal with a wide range of practical applications quite different from the classical application mentioned above. One such diverse application of the theory in an unlikely domain will be discussed in this paper.
Keywords: Credibility Theory, Prior Distribution, Likelihood Function, Posterior Distribution, Loss Function, Bayesian Approach
View PDF