Algorithmic paradigms in Big Data

Abstract

The gigantic increase in the volume of data that needs to be stored, transmitted and processed in recent years has led to the emergence of a new category of algorithms designed specifically for what is called Big Data. Today's information is too large, too diverse and requires real-time transmission to be manipulated by classical techniques. This has led to a complete rethinking of algorithmic paradigms. In this paper we present the MapReduce paradigms, streaming algorithms, approximate structures, algorithms for large graphs, dimensionality reduction, distributed machine learning. For a better understanding we have added comparisons, intuitive explanations, mathematical formulas and pseudocode.