Approximate Reasoning
James J. Buckley, Esfandiar Eslami · 2002
A method of processing information (data) through fuzzy rules is called approximate reasoning. If we have only one fuzzy rule like “if size is big, then speed is slow”, and we are given a (fuzzy) value for size, then approximate reasoning gives us a method of computing a conclusion about speed. The terms “big”, “slow” and the data for “size” are all represented as fuzzy sets. The single rule case is discussed in the next section and multiple fuzzy rules are studied in the third section. Also in the third section of this chapter we look at two methods of evaluating a block of fuzzy rules: (1) FITA, or first infer and then aggregate; and (2) FATI, or first aggregate and then infer. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.