Biomedical pattern classification using an optimized fuzzy adaptive logic network

Our research focuses on the classification of biomedical data using optimized fuzzy adaptive logic networks (FALN). The FALN exploits the advantages of geometric representations of data and fuzzy logic-oriented operations on those representations. The construction of an optimized FALN involves the selection of a heterogeneous structure of varied learning architectures (for example, multilayer perceptrons (MLP) and radial basis functions (RBF)) for the geometry-based processing subsystem. Parametric gradient-based learning is adequate for optimizing an individual MLP, since the backpropagation of the error may be described using a differentiate function. However, gradient-based learning cannot be used for the learning of an optimal heterogeneous structure. One possible solution to this structure optimization problem is to use a genetic algorithm (GA) to find an optimal combination of MLPs and RBFs. This paper presents a solution to the FALN structure optimization problem using a GA approach. The performance of the optimized FALN is demonstrated by classifying a complex biomedical dataset comprising of 186 infrared spectra of synovial joint fluid