Big Data Clustering

Abstract

Clustering algorithms group data items based on clearly defined similarity between the items aiming to minimize the intracluster differences and maximize the intercluster distances. A wealth of efficient and good quality clustering algorithms are already available for traditional data, but there are challenges for applying them to big data due to the overwhelming volume and complexities of such data. Data volume is getting bigger at an incredible pace due to growing access to Internet, social media, mobile devices, and technological innovations, and improving clustering algorithms, their computational cost and scalability have been the focus of much of the research in this area. This article provides an introduction to the characteristics of big data, and an overview of available algorithms and the current improvement trend of clustering algorithms for dealing with the challenges of big data.

Min‐max kurtosis stratum mean: An improved K‐means cluster initialization approach for microarray gene clustering on multidimensional big data

“…Multiple machine‐based clustering environments distribute the workload across multiple servers and commodity machines by parallel execution such as map‐reduced or peer‐to‐peer network variations. A single machine environment uses single server resources through batch processing and enables the addition of additional resources as necessary 2,23,24 …”

Maxmin Data Range Heuristic-Based Initial Centroid Method of Partitional Clustering for Big Data Mining

“…Classical clustering algorithms are facing various challenges due to data volume, variety, and velocity. The data volume is defining the computational cost, speed, efficiency, and scalability challenges of the classical clustering algorithms (Khondoker, 2018;Maheswari & Ramakrishnan, 2019). Big data clustering focuses on scale-up, speed-up, optimizing computation costs, and resources without the effect of cluster quality.…”

“…Big data clustering focuses on scale-up, speed-up, optimizing computation costs, and resources without the effect of cluster quality. The design of the big data clustering is dependent upon the single-machine and multiple-machine execution environment (Khondoker, 2018).…”

Euclidean distance stratified random sampling based clustering model for big data mining

“…Data mining algorithms are required to improve upon their computational cost, speed, scalability, flexibility, and efficiency according to the essential characteristics of big data 7 . Extraction of appropriate hidden predictive information, patterns and relations from the heterogeneous large‐scale dataset is big data mining, 8 which requires higher transparency for volume, variety and velocity because large‐scale data contains valuable knowledge and information 9 .…”

“…Contributions in the past 7,15,16,19,20 addressed cluster creation techniques under single and multiple machine execution environments of big data mining. Clustering techniques for big data mining are categorized into divide‐and‐conquer, parallel, center reduction, efficient nearest neighbor (NN) search, sampling, dimension reduction, incremental, and condensation methods.…”