A Comprehensive Study of Big Data Machine Learning Approaches and Challenges

Abstract

Big data is spreading its span in almost every walk of science and engineering. Both public and private sector enterprises have been collecting and deploying enormous amount of domain-specific information to gain insights about areas like security, marketing, forecasting, fraud-detection, strategic planning etc.. This big data potential is unquestionably noteworthy; but to explore it fully and sensibly it requires new ideas and original learning techniques to address challenges associated with it. With the universe being getting more knowledge-based and computerized, an enormous range of applications shows interest in machine learning (ML) techniques. Machine learning is one of the most sought after field to handle big data challenge. With this paper we endow with a literature analysis related to the up-to-the-minute progress in researches on big data processing deploying Machine Learning as an analytical tool. We will review machine learning techniques with a focus on the promising learning methods like transfer learning, active learning, deep learning, representation learning, distributed, kernel-based learning and parallel learning. Also we will be reviewing the challenges in big data machine learning.