Concept data analysis. Theory and applications. With a foreword by Peter Eklund. (English) Zbl 1083.68117

Hoboken, NJ: Wiley (ISBN 0-470-85055-8/hbk). xviii, 201 p. (2004).
The book presents theoretical and algorithmic foundations of the field of concept data analysis and processing, and discusses various applications of it in information retrieval, data visualization, and machine learning. The book consists of two parts, theory and applications. The first part tackles the theoretical foundations of the field. It has two chapters. In Chapter 1, the authors present the basic notions of ordered sets and intersection structures, contexts, concepts and concept lattices. Examples used are easy to understand and follow. Chapter 2 presents a set of algorithms for solving specific tasks, such as construction and maintenance of concept lattices, their visualization on a computer screen, and their transformation to account for additional existing knowledge. The algorithms for lattice construction are distinguished according to the generated output. The algorithms for lattice update are divided into two classes, namely algorithms for incremental construction of a concept lattice as new objects are added, and algorithms for updating lattices in response to general variations of the context table. The second part of the book illustrates how concept data analysis is utilized in a broad range of text and data processing applications. It consists of three chapters. Chapter 3 discusses the use of concept lattices for two fundamental information retrieval tasks, namely interactive query modification and automatic text ranking. Chapter 4 deals with applications of concept lattices in text mining. It presents two case studies. The first one explores mining the content of portions of the ACM Digital Library, and presents a lattice-based framework for querying, browsing, and bounding it to discover information that is hard to acquire by means of conventional searching techniques. The second study is concerned with mining Web retrieval results. A system, called CREDO, is presented to illustrate visualization and navigation of the content of the documents retrieved by a Web search engine. In Chapter 5, the authors discuss rule mining, and more specifically inference of implications, inference of functional dependencies, extraction of association rules, and induction of classification rules.


68T30 Knowledge representation
68P20 Information storage and retrieval of data
68T05 Learning and adaptive systems in artificial intelligence
68-02 Research exposition (monographs, survey articles) pertaining to computer science
06B99 Lattices