Big data and social networks
A personal recommender system based on an unweighted graph
Nowadays, many organizations, companies and researchers need to deal with big datasets in the order of terabytes or even petabytes. A popular data processing engine for handling such big datasets in an efficient way is Hadoop MapReduce. These big datasets are often represented as graphs by many systems of current interest (i.e. Internet, social networks).
A key feature of these systems is provided by personalized recommender systems for information retrieval and content discovery in today’s information rich environment. Usually, modern recommendation systems use complex techniques to provide advices to each user.
In this paper we explain our implementation of different way to provide recommendations to users of a network based on an unweighted graph, using an Hadoop iterative MapReduce approach for the implementation.
