Optimizing a Personalized Movie Recommendation System with Support Vector Machine and Content-Based Filtering

Personalized movie recommendation systems have become increasingly popular in recent years.Support Vector Machine (SVM) and Content-Based Filtering (CBF) are popular techniques for building such systems.Cold start is a problem in any recommendation system requiring a specific method to address.This paper discusses the use of SVM and CBF in building a personalized movie recommendation system.The cold start problem is addressed by jump-starting the recommendation system.The recommendation system and web services are developed as a web-based application as proof of concept and feasibility.The application interface for recommendation is in Top-N List, and the technology used to provide recommendations is Attribute-Based Recommendation with Organic Navigation.Movie datasets are obtained from the Movie Database (TMDB).The hyperparameter tuning process yields the best accuracy of 88.5% with RBF kernel function (C=1 and gamma=10) on 250 movie datasets.Further analysis shows that movie trailers and posters are two attributes that affect user preferences on liking/disliking a movie.