Intelligent Systems and Financial Forecasting

Rating:
Summary: Good exposition of AI and a financial application
Review: This book contains a number of new and interesting insights into the application of neural networks and genetic algorithms for finance. The idea of using "network regression pruning" looks extremely powerful and it would be very interesting to see this compared with other network pruning techniques. An aspect also worth highlighting is the detailed treatment of the financial experiments which appear to have been conducted in a rigorous and careful manner. I have attended many seminars and talks in which the experimentation for financial applications has been extremely poor. In many instances there is a failure to establish objective test criteria. In this work a single test run on ten years' worth of data is conducted and the algorithm performs well, contradicting all forms of the efficient market hypothesis. The level of detail also allows the reader to reconstruct the algorithm and methods for their own use. Other books on this subject should be as clear and as open.

Rating:
Summary: A good book for researchers.
Review: This is a good book, but potential buyers should realize that it is aimed at researchers, not traders. It would take considerable work to rediscover the successful neural net discussed in the book, and there is no guarantee that the net would do well in real trading because the author does not seem to have taken slippage and commissions into account.