Machine Learning in Big Data handling: Challenges and Prospects

Data is the buzzword people use these days. It is widely used by researchers to end users across the globe, any kind of information flow happens only through data. During last few decades' human evolution ends up in using huge number of software-based operations which generate humongous data as output. Evolution of newer technologies, tools, concepts also generate huge amount of data, consider a social media app which is used by a million people, it produces petabytes of data per day. Companies across the globe focus on user friendliness and enhancing user experience through utilization of certain process which requires processing of huge data. Considering this research across the globe had taken numerous efforts to produce specific tools and techniques to process voluminous data i.e. BD “Big Data”, one such process or concept is ML “Machine Learning”. ML came as a saviour owing to its versatility, usability and good interface, connectivity, interoperability, and other features, rapidly replaced other tools used for BD processing. In this work researchers had made an attempt to analyse several ML algorithms utilized in processing BD, its usability, concept involved, prevailing bottlenecks, applicability and versatility. Not all algorithms are viable and usable for all cases, hence proper selection of algorithm should be the task of the user, which will be derived based upon experience and expertise. ML in spite of the various issues is rapidly and efficiently replacing obsolete tools used for BD processing and soon will become monopoly.