Expert Systems: Principles and Programming
Joseph C. Giarratano and Gary D. Riley
Fourth Edition, Course Technology, Boston, MA, 2004
856 pages, $131.95
Expert systems is one of the most successful, practical, and recognizable subsets of classical Artificial Intelligence. The ability to supply decisions or descision-making support in a specific domain has seen a vast application of expert systems in various fields such as healthcare, military, business, accounting, production, video games, and human resources. The theoretical and practical knowledge of expert systems is indispensable for a computer science, management information science or software engineering student.
In its fourth edition, Expert Systems: Principles and Programming, as the title suggests, aims to be used as a complete textbook on this topic. The authors are respected authorities on expert systems, and were involved in the development of the popular CLIPS expert system tool which is dealt with thoroughly in this book. The book itself is divided into two major sections: the first six chapters deal with the theory of expert systems, the rationale behind their historical development and the current state of research; the next six sections are an introduction to CLIPS and how to use it to develop practical applications. The clear division between theory and practice serves to guide the student in choosing specific topics throughout the book. Each chapter ends with a set of problems and a useful bibliography. Appendix G provides a comprehensive list of software resources and will prove to be a very valuable asset to the student interested in exploring the practical aspects of expert systems as well as those who will be developing commercial applications incorporating expert systems.
The first six chapters provide an indepth intoduction to expert systems, and deal with the representation of knowledge and methods of inference and reasoning. Each of these topics are introduced from scratch—for example, when dealing with knowledge representation, logic is described from its very basics, starting from propositional logic. The chapters on reasoning under uncertainty and inexact reasoning are very well constructed and are the most attractive feature of the book for me. Topics such as fuzzy logic and Dempster-Shafer theory are explained in good detail along with their practical significance. There is an objective flow through the chapters which helps to tie in the concepts and give an understanding on the progress of topics. Also, the disadvantages and pitfalls behind expert systems in general, and specific topics are well documented.
The second section focusing on the CLIPS expert system tool is meant to be an aid in understanding and reinforcing the concepts of the first section, but does not require a thorough reading of the latter. CLIPS, developed in part by the authors at NASA, has become quite popular as a tool for studying expert system in many university courses as well as being used in several commercial and industrial applications. The expert system programming in CLIPS does not require much experience with programming and can be picked up rapidly thanks to its simple syntax and the helpful examples in the book. A new feature of the fourth edition is the introduction of COOL, the CLIPS Object-Oriented Language, which allows expert systems programmers to develop their systems in an object-oriented environment. The section is supplemented by a CD containing Windows and MacOS executables for CLIPS and reference guides—all of which can be downloaded from the Internet as well.
Although the CLIPS examples deal with some problems of uncertainty in reasoning, there is no mention of using fuzzy logic or Dempster-Shafer theory in a practical setting—a serious disadvantage to the effectiveness of the book, considering the availability of tools such as FuzzyClips. The book itself is expensive—on the wrong side of a hundred dollars, but that seems to be the trend for academic books these days and shouldn't be a deterrent to anyone interested in the subject considering it has become one of the standard textbooks on expert systems.
Department of Computer Science
Southern Illinois University
Carbondale, IL 62901, USA