The trapezoidal decision shape is used to classify data that MUST be a set member for some period -- i.e. if the data represented people's height, you may wish to have a trapezoidal set that has a linear portion for "medium" height people, such as 5' to 5'7". This eliminates Fuzziness in a region, as in this case a 5'3" person would have 100% membership in the 'medium' data set. The trapezoidal waveshape does not need to be symmetrical: clicking the picture below the "trapezoidal" option button in the Waveshape Setup window will cycle through the most common shapes. Another example of trapezoidal use would be a set for automobile speed. The linear (flat) portion of the shape could be butted up at the zero (0) point and continue to 15mph, at which point it slowly drops down to 0% membership at 25mph. This would 'force' 0 - 15 mph as SLOW.
