Performance Analysis of Big Data Based Mining and Machine Learning Algorithms: A Review

Abstract

Recent day’s data is changing at an unprecedented rate in the world of data that will affect on our way to live. Challenges of big data addressing for the capturing, managing, analyzing, storing, and visualizing the big data. By these features, one can imagine the capability of big data in today’s life – but certain questions may arise in future that how it will be capable? Because of the increase in data day by day and mostly advances in analytics technology. Simultaneously we also want to improve our analytical computing in terms of performance evaluation and optimization of the QOS parameter for instant processing. For research and industry, in coming years, the ability to leverage Big Data is critically increasing.  In data market, data becomes an important strategic asset for survival of most of the industries in the data market. Only these industries are in the race and for those, who ignores the revolution risk are left behind and will not be able compete in the data market. The objectives of this research focus on the optimization of the QOS parameters such as accuracy, load, speed, security, trustworthiness of data by using the greedy approach of Artificial Intelligence and Machine Learning. This study comprises numerous categories of optimizing algorithms, which are referred and compared with resulting parameters to reach the specified goal. The optimized outcomes will help to design the resultant algorithm that will be capable to process any real-time data instantly. To improve the big data performance, good analysis is supported by machine learning methods. Hadoop simulator like YARN Scheduler Load Simulator (SLS) is used to solve such kind of task or problems.