A fuzzy classifier using continuous automata
Jerry Jose Zachariah, Abdul Nizar M. · 2015
Classification is a common task in pattern matching and machine learning. It starts with a training set of input vectors whose classes are already known and builds a classifier mode that maps a new input vector onto the correct class. Fuzzy logic is a powerful tool that can be used to efficiently interpret noisy and imprecise data. Fuzzy classification systems are rule-based fuzzy systems which can perform a classification task based on fuzzy logic. Continuous (cellular) automata is a parallel computational model that consists of simple interconnected units called cells having continuous states. In this paper, we build a fuzzy classification model using continuous automata. As continuous automata is inherently parallel, a fuzzy classification system based on that can leverage performance in massively parallel systems. Our experiments show that the new method has a higher degree of accuracy than existing schemes.