Transitive fuzzy logic inference for the nearest pattern search

Abstract

The fidelity based transitive fuzzy-logic inference is utilized to accelerate the nearest neighbour pattern search in the hierarchically categorized filed patterns. The extraction method of inference rules and the transitive inference method of categories are discussed. Inference hardware configuration is shown and performance is estimated. The effect of an intuitive leap on flexible selecting and filtering of categories by the transitive fuzzy-logic inference are cleared.<>