The Case for Defuzzification
by Kim Pederson…….
I stumbled across and over an odd phrase in a work assignment a few days ago: Basic Defuzzification Distribution Method. What the…? I knew it was a statistical things from the context but beyond that I had no clue. To make things more clear, or to defuzzify if you will, defuzzification refers to “the process of producing a quantifiable result in fuzzy logic.” Hmm. What we have here is a failure to defuzzify. Forging ahead stoically, however, fuzzy logic is, well, you don’t want to know the official definition, believe me.
Put more in lay terms, the story goes like this. Classic logic only permits conclusions that are true or false, as in 1+1=2 (in Base Ten of course). However, human and animals, Wikipedia tells me as if I didn’t know, operate using fuzzy evaluations in many everyday situations. When you throw a ball to someone standing 20 feet away, for example, you do not calculate “exact values for the object weight, density, distance, direction, container height and width, and air resistance” to determine how hard to throw the ball to get it to the one waiting to catch it. No instead, you instinctively apply quick “fuzzy” estimates based on prior experience to make the throw. To give you an idea of how well this works (not!) if you are lacking in the experience department, just think about all those incredibly lame efforts to throw out the first pitch at baseball games.
Fuzzy logic is useful, so I’m told, when you have to come to a conclusion from variable information– for example, deciding what color something is after asking 20 people to tell you what they think it is and the answer, instead of being a plain and simple “red,” ranges from fire engine red to Ferrari red to incarnadine to, believe it or not, Fuzzy Wuzzy. In such cases, “the truth appears as the result of reasoning from inexact or partial knowledge.”
That revelation brings to mind another fuzzidiom: “fuzzy math.” On the one hand, fuzzy math is a bona fide branch of mathematics that employs fuzzy set theory and fuzzy logic. On the other, the phrase crops up when one politician (thank you, George W.) dismisses the statistics quoted by another politician as being doubtful or inaccurate (so sad, Al). George W. never explained his use of the term, though, as a New York Times writer once pointed out:
But I was chagrined when Mr. Bush left it at that. Although he repeated the phrase several times, he never discussed specific problems with Mr. Gore’s reasoning or his figures, nor did he offer any alternatives. I gradually came to understand that by ”fuzzy math” Mr. Bush meant, ”Math is confusing and fuzzy, so ignore it.”
The upshot of all this is that we are surrounded by fuzziness; we have, to put it bluntly, been fuzzombified in many respects, whether its politics, advertising, world-national-local news, Miley Cyrus, whatever. We are in dire need of a universal defuzzicator, in the same way that Dr. Who’s TARDIS and the babblefish are universal translators. If everyone had one of these on hand, or better yet permanently implanted in our frontal lobes, then maybe the truth, such as it ever is, would appear from something other than “inexact and partial knowledge.”