An effective and efficient clustering based on K-means using mapreduce and TLBO

Abstract

A plethora of clustering methods were developed since time unknown, but these methods have failed to prove that they are flawlessly efficient and also to give an optimized result in the field it might be that, parallel programming technique like MapReduce and evolutionary methods of computation address solutions to this issue as well. We use this limitation as an advantage to combine a new efficient method for optimization, ‘Teaching Learning based Optimization (TLBO)’ and a new parallel programing technique called MapReduce to develop a new approach to provide good quality clusters. In this paper, teaching learning based optimization is collaborated along with Parallel K-means Using MapReduce. Firstly, it makes Kmeans with MapReduce to work with massive amount of data and after that it takes the advantage of global search ability of TLBO to provide a global optimal result.