Big Data Analytics with Machine Learning Tools

Abstract

Big data analytics is the current, trending hot-topic in the research community. Several tools and techniques for handling and analysing structured and unstructured data are emerging very rapidly. However, most of the tools require high expert knowledge for understanding their concepts and utilizing them. This chapter presents an in-depth overview of the various corporate and open-source tools currently being used for analysing and learning from big data. An overview of the most common platforms such as IBM, HPE, SAPRANA, Microsoft Azure and Oracle is first given. Additionally, emphasis has been laid on two open-source tools: H2O and Spark MLlib. H2O was developed by H2O.ai, a company launched in 2011 and MLIB is an open source API that is part of the Apache Software Foundation. Different classification algorithms have been applied to Mobile-Health related data using both H2O and Spark MLlib. Random Forest, Naïve Bayes, and Deep Learning algorithms have been used on the data in H2O. Random Forest, Decision Trees, and Multinomial Logistic Regression Classification algorithms have been used with the data in Spark MLlib. The illustrations demonstrate the flows for developing, training, and testing mathematical models in order to obtain insights from M-Health data using open source tools developed for handling big data.