Pattern Recognition Algorithms for Data Mining

Kybernetes

ISSN: 0368-492X

Article publication date: 1 August 2005

133

Keywords

Citation

Hutton, D.M. (2005), "Pattern Recognition Algorithms for Data Mining", Kybernetes, Vol. 34 No. 7/8, pp. 1291-1292. https://doi.org/10.1108/03684920510606055

Publisher

:

Emerald Group Publishing Limited

Copyright © 2005, Emerald Group Publishing Limited


The development of pattern recognition algorithms for data mining has advanced considerably from the days when readers researching in cybernetics and systems had to not only construct their own algorithms but also program them for their computer systems.

Now we are offered both a variety of algorithms and the software and systems. A great deal has been written and a choice of publications makes it difficult to choose which texts or manuals are worth buying.

This book does what the title says it will do. It looks at a variety of pattern recognition tasks and produces the results which are both experimental and theoretical. Some of these are those that occur on a regular basis in most research and development programmes. This text covers some of the most regularly used: condensation of data, case generation, clusters and classes, as well as rule generation and evaluation. The authors have chosen to use traditional approaches as well as hybrid paradigms to present their selected theories, methods and algorithms. They have also given special treatment to a number of methodologies, presenting strategies for handling many problems as well as offering descriptions of design procedures. They use fuzzy sets, rough sets, genetic algorithms and neural nets to classification problems for example.

Where experimental results are given they are based on real‐life data and the analysis links up with the theoretical discussion.

This book contains a wealth of detail that has been organised to interest and inform the reader. Much of the text will be very useful to anyone with data mining tasks and it does not rely on previous work although a sound mathematical background will obviously enhance the understanding of the authors' chosen methodologies and general approaches to pattern recognition.

Related articles