Recommender systems are widely used commercially and academically that involve predicting user response to certain options. Since the recommendation algorithms are computational intensive and usually operate on large datasets, such problems can be efficiently dealt with cloud architecture. The algorithms could be broadly classified as Content-based and Collaborative-based. Content-based systems recommend items based on the properties of the items whereas Collaborative filtering systems recommend items based on similarity measures between users and/or items. This project aims at building a recommendation engine using Amazon AWS technologies (Hadoop, HDFS, Map-Reduce) with focus on collaborative filtering (CF) algorithms and performs comparative studies using Mahout library. Currently, this project emphasizes on predicting user-movie rating and recommending movies to the users based on user ratings using Netflix dataset.
sivasrinivas / recflix Goto Github PK
View Code? Open in Web Editor NEWThis project aimed at generating recommendations based on item based and user based colloborative filtering and comparing them agains Mahout version recommendations for performance analysis.