Towards a Hybrid Recommendation System On Apache Spark

Abstract

Now a day's success of internet business depends on its capability in providing personalized experiences for the users. In the era of SMAC, Social, Mobile, Analytics & Cloud the data is dynamic. But as the digital data is exponentially increasing users are having a deluge of options for services and commodities. Recommender Systems help users in overcoming the paradox of alternatives. This paper precis different approaches for Content based, Collaborative and hybrid recommendation systems to handle the usual problems of cold start and data sparsity. To generate accurate recommendation a hybrid frame work is proposed on the score. Experiments on movie lens dataset justify that the model proposed bring out better recommendations than the standard methods.