Big Data Formation, Reduction, and Its Impact on Sampling: A survey

Abstract

The emergence of data in recent years, characterized by the "6Vs" (Volume, Velocity, Variety, Veracity, Value, and Variability), has started the era of big data. While this data holds great potential for uncovering valuable insights and knowledge, its size presents significant challenges for analysis. This paper explores two critical big data reduction techniques: feature selection and sampling. Feature selection focuses on identifying and eliminating irrelevant or redundant features, reducing data dimensionality. Sampling, on the other hand, selects a representative subset of data points for analysis. We compare and contrast these techniques, highlighting their strengths and weaknesses. The paper explores when each approach is most suitable and suggest the potential benefits of combining them for even more efficient big data analysis.