Abstract
While many techniques exist to classify data possessing straightforward characteristics, they tend to fail when dealing with the "curse of dimensionality". This condition, in which the ratio of features to samples is very large, is prevalent in many complex, voluminous biomedical datasets acquired using current spectroscopic modalities. We present a novel classification method using an adaptive network of fuzzy logic connectives to combine class boundaries generated by sets of linear discriminant functions. We empirically demonstrate the effectiveness of this method using a benchmark linear discriminant analysis approach with feature averaging.