Machine Learning Techniques for Dimensionality Reduction in Big Data

Abstract

The incredible growth in information assortment and data-storing abilities during the previous years had prompted data over-burden in many areas. Scientists works in different areas such as scientific laboratory, construction field, cosmic science, geographical information system, share markets, and spatial fields due to this the volume of data are increasing periodically. Because of its elaborating datastructure, it leads to huge difficulties in the analysis of data engineering. Higher dimension data leads to numerous numerical difficulties with certain chances that provide us to propose innovative methods/solutions to overcome these issues. The major issue in higher dimension data are is to distinguish the relationship between useful and non-useful data. Meanwhile, some models were proposed by various researchers to handle the dimensionality problem by providing better accuracy in terms of reduction in unwanted data, but it still faces a lot of deviation. In this survey article, we had analyzed various methods for the data reduction process. We had found that dimensionality reduction methods are providing a good solution to reducing the unwanted volumes in big data. The role of a machine learning algorithm and some feature techniques are very much supportive in this task. Thus this article will be very useful for researchers in the data analytical field to make their work efficient.