Performance evaluation of ensemble based recommendation system using biased matrix factorization and content based filtering
Authors
Deogirikar, Jayesh
Issue Date
2020
Degree
MSc in Data Analytics
Publisher
Dublin Business School
Rights holder
Rights
Items in eSource are protected by copyright. Previously published items are made available in accordance with the copyright policy of the publisher/copyright holder.
Abstract
Boosting product sales is the primary objective of any Recommender System and it is achieved by providing users a personalized recommendation from the world of information density and product overload. The suggestions given by RS support users in their decision-making process such as which item to buy from thousands of items, which music to listen to from millions of soundtracks or which news feed to read rst. Recommender System can be implemented by using Collaborative or Content-based ltering techniques and each of them has its own merits and demer- its. Aiming at this concern, this research proposed an ensemble model combining Item-KNN, Biased Matrix factorization and user pro le content-based ltering. The result for comparative analysis on MovieLens data shows that the ensemble model outperforms each model when implemented individually and ensemble model without BMF giving an accuracy score of 0.818 stars as RMSE and 0.629 stars as MAE.