“…4 and the parameters of the linear functions of (4) in fourth and fifth layers. To design FNS classifier, the training of the membership functions parameters c ij (t) and ij (t) (i=1,..,m, j=1,..,r) in the premise part and parameter values of the w jk (t), a ij (t), b j (t) (i=1,..,m, j=1,..,r, k=1,..,n) in consequent part is carried out [28,31]. In the paper, we applied fuzzy clustering and gradient algorithms for the update of FNS parameters [33].…”