Book Review: Prices and Knowledge: A Market-Process Perspective

designer" for "knowledge engineer" and "marketing information system" for "expert system," he or she would quickly get a sense of deja vu, because the problems currently faced by the proponents of expert systems are very similar to those faced by the proponents of integrated management information systems. Have we made much progress?) Many pragmatic problems have to be faced in designing an expert system; some of the more important ones and recommendations for their resolution are presented in Chapter 4. For example, one of the first problems is to summarize and edit the knowledge base. The recommendation for the resolution of this problem is to use decision trees and flow charts. Finally, in Chapter 5, the testing and validation of expert systems is discussed. Testing involves the use of multiple experts and, where possible, more traditionally statistical techniques (actually for prediction as well as for testing and validation of the expert system). The Appendix provides a list (names, prices, comments on usability, etc.) of shells and languages available for the design of PC-based expert systems. As mentioned earlier, the targeted reader of this book is not the typical reader of JMR. If a reader wanted only to learn the fundamentals of expert systems, the rather simple book by Van Hom (1986) might be more appropriate than this book. If he or she was interested in marketing-type applications of expert systems, current journal articles, e.g., the article on forecasting by Collopy and Armstrong (1992) might be more appropriate. In summary, it appears this book has missed a market niche that it wasn't even aiming for anyway.