A multilayer perceptron network for the diagnosis of low back pain

A description is given of the training of a multilayer perceptron (MLP) network for the diagnosis of low back pain and sciatica. The performance of the system is compared with that of three groups of doctors and also with another computer program based on fuzzy logic ideas. Although clinical data on larger numbers of patients would be necessary before this neural-network approach could be fully validated, the initial results are very promising. For the diagnostic category which it is most crucial to get right (because surgery may be required urgently), the MLP produces the correct classification more often than any of the three groups of doctors or the fuzzy logic system. The mean diagnostic accuracy over all possible diagnoses is also higher than that of the doctors, but slightly worse than the fuzzy logic system.<>