8.10 Review
■
337
If at least one rule was executed, the outcome value (dDVEstimate)is
generated by defuzzification. The value is a weighted average computed
by dividing the sum of the scaled center valued by the sum of the individual
rule degrees. The compatibility index (degree of evidence in the model’s
solution) is calculated as the average degree.
if(iRulesFired>0)
{
fCIX=(float)(dSum_of_Degrees/(double)iRulesFired);
dDVEstimate=(dSum_of_Centers/dSum_of_Degrees);
if(fCIX<fMinCIX)
iAvgRsltBelowCIX++;
fTotCIX+=fCIX;
}
The technical overview provides an introduction to the underlying data
structures and code models used in the rule induction engine, although
naturally it does not cover all procedures and structures.
8.10 Review
Rule discovery forms the core ...