Parallel and Scalable Map Reduce and Pipeline Tree Classifiers for Massive Dataset Using Map Reduce and Data Flow Pipeline

One of the important research areas in today's scenario is classification of Big Data. While there are a lot of traditional classification methods, extending them to Big Data is quite challenging. Decision Tree Classifier is one of the effective traditional classification techniques. The combination of Hadoop and Map Reduce has been adapted by many researchers both commercially and academically to process Big Data. Of late, Google cloud dataflow paradigm has sneaked into the Big Data scenario that augments the earlier systems with stream processing. This paper presents two algorithms based on Map Reduce and Google cloud data flow for implementing decision trees for classification is presented. The performances of both algorithms on various parameters have been compared and presented.