Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Information granularity, big data, and computational intelligence
Pedrycz W., Chen S., Springer Publishing Company, Incorporated, New York, NY, 2014. 444 pp. Type: Book (978-3-319082-53-0)
Date Reviewed: Aug 12 2015

Big data (BD) is the hottest topic in current research on data science and engineering. In a nutshell, BD refers to data collections that challenge present technologies and methodologies in terms of the four Vs: volume of data, velocity of data production, variety of data representations, and veracity of generated data. BD is generated both by devices (for example, sensors, artificial agents, and so on) and people (for example, through social networks, Internet transactions, and so on), and is the principal source of the exponential data growth we are facing since the spread of the Internet at a large scale. The interest in extracting some value from BD is huge; therefore, notwithstanding some authoritative yet skeptical positions on the possibility of mining useful knowledge from BD [1], research on BD is progressively increasing.

This is the general context in which we can put this book. In more specific terms, the editors focused their attention on two paradigms for approaching BD problems: granular computing and computational intelligence. Granular computing enables the representation and processing of information granules, that is, clumps of objects drawn together by some relation that allows viewing the objects as a whole and that may be processed as such. Fuzzy sets, rough sets, intervals, and quotient spaces are prominent examples of mathematical enabling tools for granular computing. Computational intelligence embraces all methodologies that draw inspiration from biological and linguistic systems to design intelligent computing solutions. Computational intelligence includes granular computing (because human beings reason in terms of information granules, or at least this is a widely shared credo) as well as neural networks (in all their flavors) and evolutionary computation methods. Both granular computing and computational intelligence are promising tools for mining BD because they try to mimic the ways humans and nature process information in complex environments.

The book consists of 21 chapters and is structured in three parts. The first, largest part includes 12 chapters and is centered on methodology. A number of methods are presented with focus on how to tackle BD problems (although not all chapters center this objective, unfortunately). Some chapters describe very efficient data structures to enable the application of data mining methods, possibly extended to deal with efficiency issues, like k-nearest neighbor (k-NN), frequent pattern mining, decision trees and rules, and graph analysis. Other chapters focus on granular computing and information granulation, like fuzzy regression, quotient space analysis, information granulation of user-generated data, and latent semantic indexing.

The second part of the book consists of only three chapters, which basically present some architectures for BD. The first chapter is very general and illustrates the cloud computing architecture and its role in developing business intelligence applications that deal with BD. The remaining two chapters are more specialized: one presents an architecture for storing and querying big spatial data; the other one describes a framework based on fuzzy cognitive maps to manage fuzzy ontologies generated by web data.

The last part of the book is composed of six chapters, each presenting case studies on customer relationship management (CRM), economics, finance, weather forecast, and the environment.

What I missed most in this book is, well, big data. Almost all chapters start by extolling the challenges of BD, but end with presenting experiments on data that cannot be intended as really big, and could be tackled by standard data mining methods. Disappointingly, this is also true for the chapters in the third part, which are intended to present case studies. Only one chapter (in the first part of the book) includes an experiment on seemingly real big data (intelligence data and bibliography data), although the used data is not described in terms of the four Vs (see above). Overall, my general impression of the book is that it is a collection of position papers, where interesting ideas have been put forward but have not been tested in really challenging scenarios such as those encountered in real-world BD applications. This makes the book interesting for researchers who want to plunge into the wide world of BD analytics, but it is far from being a reference text on the subject.

Reviewer:  Corrado Mencar Review #: CR143683 (1511-0943)
1) Gomes, L. Machine-learning maestro Michael Jordan on the delusions of big data and other huge engineering efforts. IEEE Spectrum, Oct. 20, 2014. http://bit.ly/1Ih0YPz.
Bookmark and Share
  Reviewer Selected
Featured Reviewer
 
 
Applications And Expert Systems (I.2.1 )
 
 
Fuzzy Set (I.5.1 ... )
 
 
Content Analysis And Indexing (H.3.1 )
 
Would you recommend this review?
yes
no
Other reviews under "Applications And Expert Systems": Date
Control structures in expert systems
Laurent J. Technology and Science of Informatics 3(3): 147-162, 1984. Type: Article
May 1 1985
Knowledge-based expert systems
Hayes-Roth F. Computer 17(10): 263-273, 1984. Type: Article
Jun 1 1985
Expert systems and fuzzy systems
Negoita C., Benjamin-Cummings Publ. Co., Inc., Redwood City, CA, 1985. Type: Book (9789780805368406)
Aug 1 1985
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy