Modular Fuzzy Hyperline Segment Neural Network

This paper describes Modular Fuzzy Hyperline Segment Neural Network (MFHLSNN) with its learning algorithm, which is an extension of Fuzzy Hyperline Segment Neural Network (FHLSNN) proposed by Kulkarni and Sontakke. The MFHLSNN offers higher degree of parallelism. Each module in MFHLSNN is exposed to the patterns of only one class and trained without overlap test and removal, unlike in FHSNN, leading to reduction in training time. Hence, each module captures peculiarity of only one particular class and due to decrease in training time the algorithm can be used for voluminous realistic database, where new patterns can be added on fly. The MFHLSNN is found superior than FHLSNN in terms of generalization and training time with equivalent testing time.