Movie Master: Hybrid Movie Recommendation

摘要

A recommendation system provides an individual with personalized service. This paper describes our research conducted to develop and implement a Movie Recommendation engine in the form of a Web Application using two simple approaches: (1) Non-Personalized Recommendation, (2) Content based recommendation techniques using a machine-learning algorithm. The former is achieved by Bayesian Estimation and the latter is derived based on Term Frequency and Inverse Rating Frequency(TF-IRF) Approach coupled with the Cosine Similarity Measuring Technique. Our results indicate that the proposed approach Bayesian Estimation and TF-IRF approach is efficient in terms of calculating the prediction and recommendation factor for a movie with a minimum webpage loading time, when compared to the existing methods such as Aggregate Opinion Mining and Product Association.