Design of friend recommender system using apache hadoop
Lakshay Nagpal, Nikhil Khurana · 2017
Modern era applications involve social networks to form a great platform to share information and interaction among users. A friend recommender system is critical to expand the networks by actively recommending new friends. It has been observed by concrete studies that a simple system with large amounts of data is much better than a complex system with little data. As giving recommendations is a data intensive task, we have used an open source system provided by Apache Software Foundations to analyze such a large amount of data in a relatively small amount of time. In this paper, the improved MapReduce based data processing are proposed and implemented using Hadoop framework. In our system, recommendations are based on number of mutual friends in the given network. Our system focuses on the friends with whom people have the most friends in common. The scalability and efficiency of Hadoop makes it an ideal platform to implement recommender algorithms for large-scale social network graphs.