Novel Fuzzy Inference System (FIS) Analysis and Design Based on Lattice Theory

摘要: We introduce novel (set- and lattice-theoretic) perspectives tools for the analysis design of fuzzy inference systems (FISs). present an FIS, including both fuzzification defuzzification, as a device implementing function f: RNrarr RM. The family FIS functions has cardinality aleph2=2aleph1, where aleph1 is set R real numbers. Hence much larger than polynomials, neural networks, etc.; furthermore capacity local generalization. A formulation in context lattice theory allows us to define F* interval numbers (FINs), which includes (fuzzy) intervals. metric dK on F*, can tunable nonlinearities. based advantages such as: alleviation curse dimensionality problem potential improved computer memory utilization. new classifier, namely granular self-organizing map (grSOM), we apply industrial fertilizer modeling application