Scalable recommender system based on MapReduce framework
Rohit Rohit, Anil Kumar Singh · 2017 IEEE International Conference on Power, Control, Signals and Instrumentation Engineering (ICPCSI) · 2017
The data collection of items and its users are growing day by day over the internet. Recommender Systems (RSs) help the user to find the items of their choices. It is a big challenge for RSs to identify the right choice of the items for a user from the ocean of large data set. Therefore, the scalability of RSs is an important demand for today's application. MapReduce framework provides a solution for such demand of scalability. Hadoop is an open source tool, that helps to build such applications. The performance analysis of scalable RSs needs to be verified. In the paper, a content based RS is implemented using the MapReduce framework based on Hadoop. The scalability of RS is verified in terms of parallel computation of each task of MapReduce framework and their data distribution. The resource consumption and the execution time of the algorithm are observed in the experiment. The outcome of the experiment shows the good trend for the demand of scalability of RSs.