Information Technology Reference
In-Depth Information
Fig. 18.13
Original fitting of the fuzzy model for discriminant analysis,
Syst3
vs.
Age
and
Bmi
(the first-order Takagi-Sugeno model)
Fig. 18.14
Fitting of the fuzzy model for discriminant analysis when the outputs are rounded
into integers.
Syst3
vs.
Age
and
Bmi
.
In fact, the ideas on switching regression modeling and fuzzy clusters are also
applied to the first-order Takagi-Sugeno reasoning algorithm because it generates
fuzzy rules the consequents of which are linear functions based on the data clus-
ters [19]. In this case the output fitting is nevertheless the weighted sum of these
functions (Fig. 18.18).