Training type-2 Fuzzy System by particle swarm optimization

Optimal design of adaptive type-2 neuro-fuzzy systems: A review

“…A training method for a T2FLS using PSO was presented by Al-Jaafreh and Al-Jumaily in [46]. T2FLS and PSO were utilized together, the procedure to analyse the problem was explained and finally presented a new method to optimize parameters of the primary MFs of T2FLS using PSO to improve the performance and increase the accuracy of the T2FLS.…”

Soft computing applied to the build of textile defects inspection system

“…Instead of merging the type-2 sets from the fuzzy inference of all the rules before reduction, the centre-of-sets type-reduction makes use of the centroid method to reduce the resulting type-2 sets from each rule and obtain a type-1 set y i l , y i r for each rule i. The weighted mixture of these type-1 sets is then used to obtain the final type-1 set [y l , y r ] (26) where N r is the number of rules and f i r , f i l are the firing level corresponding to y i r , y i l of rule i that can be enumerated in the interval f i , f i .…”

“…Computing system accuracy reiterated for each particle until a steady IT2FS accuracy or the maximum iteration has been obtained. The PSO individuals are evaluated by the fitness function defined by [26] where S g is the samples number of correct estimated defects, and S a is the number of all samples of each defect class. This fitness function shows advantages in terms of easy configuration and low maintenance efforts.…”