An asymmetry-similarity-measure-based neural fuzzy inference system

In this paper, a new asymmetry-similarity-measure-based neural fuzzy inference system (ASM-NFIS) is proposed. A pseudo-Gaussian membership function can provide a neural fuzzy inference system which has a higher flexibility and can approach the optimized result more accurately. An on-line self-constructing learning algorithm is proposed to automatically construct the ASM-NFIS. It consists of structure learning and parameter learning that would create adaptive fuzzy logic rules. The structure learning is based on the similarity measure of asymmetric Gaussian membership functions, and the parameter learning is based on a supervised gradient descent method. Computer simulations were conducted to illustrate the performance and applicability of the proposed model.