The importance of having efficient and effective methods for data mining and knowledge discovery (DM) is rapidly growing. This is due to the wide use of fast and affordable computing power and data storage media and also the gathering of huge amounts of data in almost all aspects of human activity and interest. While numerous methods have been developed, the focus of this book presents algorithms and applications using one popular method that has been formulated in terms of binary attributes, i.e., by Boolean functions defined on several attributes that are easily transformed into rules that can express new knowledge.
This book presents methods that deal with key data mining and knowledge discovery issues in an intuitive manner, in a natural sequence, and in a way that can be easily understood and interpreted by a wide array of experts and end users. The presentation provides a unique perspective into the essence of some fundamental DM issues, many of which come from important real life applications such as breast cancer diagnosis.
Applications and algorithms are accompanied by extensive experimental results and are presented in a way such that anyone with a minimum background in mathematics and computer science can benefit from the exposition. Rigor in mathematics and algorithmic development is not compromised and each chapter systematically offers some possible extensions for future research.
Author(s): Evangelos Triantaphyllou (auth.)
Series: Springer Optimization and Its Applications 43
Edition: 1
Publisher: Springer US
Year: 2010
Language: English
Pages: 350
Tags: Operations Research, Mathematical Programming; Mathematical Logic and Formal Languages; Operations Research/Decision Theory
Front Matter....Pages i-xxxiii
Front Matter....Pages 1-1
Introduction....Pages 3-20
Inferring a Boolean Function from Positive and Negative Examples....Pages 21-56
A Revised Branch-and-Bound Approach for Inferring a Boolean Function from Examples....Pages 57-72
Some Fast Heuristics for Inferring a Boolean Function from Examples....Pages 73-100
An Approach to Guided Learning of Boolean Functions....Pages 101-123
An Incremental Learning Algorithm for Inferring Boolean Functions....Pages 125-145
A Duality Relationship Between Boolean Functions in CNF and DNF Derivable from the Same Training Examples....Pages 147-150
The Rejectability Graph of Two Sets of Examples....Pages 151-170
Front Matter....Pages 171-171
The Reliability Issue in Data Mining: The Case of Computer-Aided Breast Cancer Diagnosis....Pages 173-190
Data Mining and Knowledge Discovery by Means of Monotone Boolean Functions....Pages 191-227
Some Application Issues of Monotone Boolean Functions....Pages 229-239
Mining of Association Rules....Pages 241-255
Data Mining of Text Documents....Pages 257-276
First Case Study: Predicting Muscle Fatigue from EMG Signals....Pages 277-287
Second Case Study: Inference of Diagnostic Rules for Breast Cancer....Pages 289-296
A Fuzzy Logic Approach to Attribute Formalization: Analysis of Lobulation for Breast Cancer Diagnosis....Pages 297-308
Conclusions....Pages 309-315
Back Matter....Pages 317-350
