Fuzzy system modeling in pharmacology: an improved algorithm (Q1855492)

The authors propose a fuzzy system modeling in pharmacology. The proposed scheme is based on an \(n\)-dimensional input space. The cluster validity of the problem demands the determination of the optimal cluster size which minimizes the modeling error. The inputs no longer classify as either significant or insignificant, but a significance degree assigns (with respect to weights in predicting quality) the output by a local search algorithm which is randomized to overcome cycles and entrapments in local optima.