A Novel Hybrid Approach for Machine Learning Based Recommended Systems

Abstract

The Recommender systems are mostly well known for their best practice in e-business sites. Classical
customized recommender algorithm incorporate item-based collaborative filtering method employed
in Amazon, matrix factorization based CF algorithm from Netflix, etc. In this article, so combining
traditional prototype algorithms and customized algorithm using Hybrid recommendation
methodology with initial recommendation list and user similarity matrix which aims to find out the
final suggestion list and evaluating and providing the best to the accuracy of the prototype and
getting the user desired list.