Introduction of R-LCS and comparative analysis with FSC and Mahalanobis-Taguchi method for Breast Cancer classification

Classification for medical diagnosis is an important problem in the field of pattern recognition. We introduce a new method for classification based on repeated analysis of information tailored to small data sets - the Rote Learning Classifier System. Using the Wisconsin Breast Cancer study, this method was compared to three other methods of classification: Mahalanobis-Taguchi Systems, Finite State Classifiers, and Neural Networks. It was found that for the given data set, the Rote Learning Classifier System outperformed the other methods of classification. This new algorithm correctly classified over 92% of the data set.