Improved Scalable Recommender System
Samin S. Ishtiaq, Muhammad Nadeem Majeed, Muazzam Maqsood, Ali Javed · The Nucleus · 2016
Recommender systems are known for their ability to recommend items which are new to the user by having some synchronization with user’s personal interest. The importance of recommender systems leads to the creation of new approaches that can produce accurate results. As data became large it results in scalability issues. In this work, we have suggested a scalable technique using different methods that work in a sequential manner. A novel centroid selection for clustering based recommender system is proposed. SVD and user representatives are used to handle scalability issues. Experiments on proposed approach with standard datasets showed great improvement in scalability and slight better accuracy