Abstract: For a type of classification problems of complex chemical patterns, in which the edges between pattern classes are rather irregular or ill defined, a new fuzzy computing method for chemical pattern classification is proposed. The new method was realized with a fuzzy neural network. An algorithm for training the fuzzy neural networks, which makes the fuzzy pattern classifier has the ability of learning from examples, is given. An example of classifying complex chemical pattern was used to verify the efficiency of the new method.

Key words: Complex chemical pattern recognition, Fuzzy pattern classification, Fuzzy neural classifier