Fuzzy Expert Systems
by J.F. Derry
Expert systems give the technical analyst a potent set of tools to dissect trading and investment
problems in short order.
For traders and investors alike, the real problem with both technical and fundamental methods is that of
laying out current after-tax dollars in the expectation of future before-tax profits based solely upon the
guidance of 20/20 hindsight generated through past experience. However investment and trading
opportunities are analyzed, investment decisions should depend upon as large a body of expertise as
possible. Such expertise can be captured in computer software in the form of an expert system, which
may contain knowledge of fundamental or technical technique or both. Although many forms of expert
systems can be (and have been) applied to securities trading and investment, for my purposes I am
concentrating on a small but usable expert system based on fuzzy logic as a demonstration project and
how to use fuzzy logic programming in everyday practice.
FUZZY LOGIC AND MEMBERSHIP FUNCTION
As originally stated by L.A. Zadeh, the concept of fuzzy logic was roughly as follows: If propositional
logic is modified to permit use of any value in the range of 0 to 1 rather than only the "crisp" binary
values of 0 and 1, then the resulting "fuzzy" logic should be able to accommodate linguistic information.