Automating the Design of Data Mining Algorithms An Evolutionary Computation Approach /

• Main Authors: Pappa, Gisele L. (Author), Freitas, Alex. (Author)
• Corporate Author: SpringerLink (Online service)
• Format: Electronic
• Language: English
• Published: Berlin, Heidelberg : Springer Berlin Heidelberg, 2010.
• Series: Natural Computing Series,
• Subjects: Computer science.; Data mining.; Artificial intelligence.; Computer Science.; Data Mining and Knowledge Discovery.; Artificial Intelligence (incl. Robotics).
• Online Access: https://ezaccess.library.uitm.edu.my/login?url=http://dx.doi.org/10.1007/978-3-642-02541-9

 Traditionally, evolutionary computing techniques have been applied in the area of data mining either to optimize the parameters of data mining algorithms or to discover knowledge or patterns in the data, i.e., to directly solve the target data mining problem. This book proposes a different goal for evolutionary algorithms in data mining: to automate the design of a data mining algorithm, rather than just optimize its parameters. The authors first offer introductory overviews on data mining, focusing on rule induction methods, and on evolutionary algorithms, focusing on genetic programming. They then examine the conventional use of evolutionary algorithms to discover classification rules or related types of knowledge structures in the data, before moving to the more ambitious objective of their research, the design of a new genetic programming system for automating the design of full rule induction algorithms. They analyze computational results from their automatically designed algorithms, which show that the machine-designed rule induction algorithms are competitive when compared with state-of-the-art human-designed algorithms. Finally the authors examine future research directions. This book will be useful for researchers and practitioners in the areas of data mining and evolutionary computation.