Intelligent Tuning of Fuzzy Controllers by Learning and Optimization - Fuzzy Modeling and Control: Theory and Applications - page 136

Fig. 7.4 Schemes for a inductive reasoning by ANFIS and b transductive reasoning by TNFIS

d i −

min d

v i = 1 − (7.2)

w here P is the number of elements in the input dat a vector, x is the input data vector,

k is each vector or datum in the training set, min d is the minimum element in the
distance vector d

= d 1 , d N q , and i

d 2 ,... = 1 , 2 ,..., N q is the index representing

the number of nearest neighbors.
Third, when the subset has been chosen and the distance weights calculated, fuzzy
rules and membership functions (with their initial parameter values) are constructed
iteratively on the basis of the closest data (i.e., each input datum) by a clustering
technique.
Evolving Clustering Method (ECM) (Kasabov and Song 2002 ) was introduced in
the previous section and the algorithm is briefly described as follows. Considering P