Computational Intelligence and Feature Selection: Rough and Fuzzy Approaches
Computational Intelligence and Feature Selection provides researchers and practitioners with both the theoretical background and fundamental methodologies behind feature selection. The monograph emphasizes techniques grounded in rough and fuzzy set theories, alongside their hybridizations.
It introduces set theory, fuzzy set theory, rough set theory, and fuzzy-rough set theory, demonstrating the mathematical rigour and empirical effectiveness of semantic-preserving feature selection across diverse real-world applications and worked examples.
- Publisher IEEE Press / Wiley
- Publication Date September 2008
- ISBN-13 978-0-470-22975-0
- Format / Pages Hardcover / 348 pp.
Monograph Structure and Topics
An outline of foundational topics covered within the text, bridging formal set-theoretic mathematics with contemporary machine learning algorithms.
Foundational Concepts
- Introduction to Feature Selection: Motivation, filter vs wrapper architectures, evaluation criteria, and search mechanisms.
- Classical Set Theory and Rough Sets: Equivalence relations, indiscernibility, lower and upper approximations, and reduct computation.
- Fuzzy Sets and Approximate Reasoning: Membership functions, fuzzy relations, t-norms, s-norms, and fuzzy implicators.
- Fuzzy-Rough Hybrids: Combining fuzzy equivalence classes with approximation spaces to process real-valued continuous data without discretization.
Innovations and Applications
- Fuzzy-Rough Feature Selection (FRFS): Degree of dependency formulations, greedy search, and heuristic evaluation metrics.
- Swarm Intelligence and Heuristic Search: Ant Colony Optimization (ACO) for minimal reduct discovery.
- Case Studies: Forensic glass evidence analysis, complex systems monitoring, and text categorization.
BibTeX Citation
@book{jensen2008computational,
title = {Computational Intelligence and Feature Selection: Rough and Fuzzy Approaches},
author = {Jensen, Richard and Shen, Qiang},
year = {2008},
publisher = {John Wiley \& Sons / IEEE Press},
isbn = {978-0-470-22975-0},
pages = {348}
}