Data driven level set modeling, machine learning modeling
Abstract
Linguistic and functional rule-based fuzzy modeling are the two dominant modeling paradigms in fuzzy systems engineering. Level set fuzzy modeling is a third rule-based paradigm whose efficiency and capability to develop data driven models and fuzzy controllers have recently been reported in the literature. It has been shown that the performance of the data driven level set models outperform adaptive neurofuzzy, neural, rule-based, and autoregressive models. This paper revisits data driven level set modeling in front of powerful machine learning models based on ensembles and kernels. The results show that the data driven level set fuzzy modeling surpass ensemble and kernel machine learning approaches as well.
