Discussion: Fuzzy Logic: Better Science? Or Better Engineering?
Russell G. Almond · Technometrics · 1995
Can statisticians afford to ignore fuzzy logic? No! Fuzzy logic appeals to many engineers in places where probabilistic solutions have not penetrated. Judging from the number of papers in recent workshops like Uncertainty in and Artificial Intelligence and Statistics, there are many traditionally statistical problems (e.g., control, classification, and decision analysis) for which researchers are using nonstatistical models. We statisticians must look at nonprobabilistic approaches to problem solving, such as fuzzy logic, if only to how we can do a better job of marketing statistics to engineers. would like to thank Laviolette, Seaman, Barrett, and Woodall (hereafter LSBW) for forcing me to go through the effort of evaluating fuzzy logic. I, like many other statisticians, have been all too tempted to say I don't understand fuzzy logic (meaning, I think fuzzy logicians are slightly potty, but haven't looked into it carefully enough to say that on the record). Laying out fuzzy and probabilistic models for the same problem side by side, as LSBW do, helps me compare the two approaches and the strengths and weaknesses of each.