A Comparative Analysis on Machine Learning Algorithms Implemented on Map Reduce Paradigm

Big data is clearly a game changer, enabling organisations to gain insights from new sources of data that was ever mined in the past. Working with such data whose size and variety is beyond the ability of typical database software to capture, store, manage and analyse. Hadoop is evolving open source framework for Big data analytics. It includes a distributed file system a parallel processing framework called Apache Map-Reduce. Map-Reduce paradigm solves many problems associated rising data volume by modelling algorithms. The machine learning field worked tremendously towards implementation of various algorithms on a Map-Reduce system. This paper presents a comparative analysis of Canopy, Min Hash and Mean Shift algorithms in a standard implementation of Hadoop in Map-Reduce distributed paradigm.