Page 12 - Big Data Analytics for Intelligent Healthcare Management

2 CHAPTER 1 BIO-INSPIRED ALGORITHMS FOR BIG DATA ANALYTICS

Dimensions of data management

Volume Variety Velocity Veracity Variability

FIG. 1.1
Dimensions of data management.

Variety of data

Cloud Machine to
Text Audio Video Social Transactional Operational service machine data

FIG. 1.2
Variety of data.

1.1.1 DIMENSIONS OF DATA MANAGEMENT
As identified from existing literature [1–6], there are five kinds of dimensions of data, which are re-
quired for effective management. Fig. 1.1 shows the dimensions of data management for big data an-
alytics: (1) volume, (2) variety, (3) velocity, (4) veracity, and (5) variability.
The Volume represents the magnitude of data in terms of data sizes (terabytes or petabytes). For
example, Facebook processes a large amount of data such as millions of photographs and videos. Va-
riety refers to heterogeneity in a dataset, which can be different types of data. Fig. 1.2 shows the variety
of data, which can be text, audio, video, social, transactional, operational, cloud service, or machine to
machine data (M2M data).
Velocity refers to the rate of production of data and analysis for processing a huge amount of data.
For example, velocity can be 250MB/minute or more [3]. Veracity refers to abnormality, noise, and
biases in data, while variability refers to the change in the rate of flow of data for generation and
analysis.
The rest of the chapter is organized as follows. In Section 1.2, we present the big data analytical
model. In Section 1.3, we propose the taxonomy of bio-inspired algorithms for big data analytics. In
Section 1.4, we analyze research gaps and present some promising directions toward future research in
this area. Finally, we summarize the findings and conclude the chapter in Section 1.5.

1.2 BIG DATA ANALYTICAL MODEL
Big data analytics is a term, which is a combination of “big data” and “deep analysis” as shown in
Fig. 1.3. Every minute, a large amount of user data is being transferred from one device to another
device, which needs high processing power to perform data mining for the extraction of useful infor-
mation from the database. Fig. 1.3 shows the model for big data analytics, which shows that an OLTP
(on-line transaction processing) system creates data (txn data). A data cube represents a big data, out of