An investigation of a linguistic perceptron in a nonlinear decision boundary problem

Abstract

We have developed a linguistic perceptron (LP) to deal with the problem in pattern recognition where inputs are uncertain. This algorithm is based on the extension principle and the decomposition theorem. Several synthetic data sets are used to illustrate the behavior of this linguistic perceptron in linearly separable, nonlinearly separable and nonseparable situations. We also compare the results from the linguistic perceptron with that from the regular perceptron. © 2006 IEEE.