An artificial neural network model called BIONET proposed in this paper, is
used for diagnosis of diseases. The diseases are classified based on a pri
ori knowledge. BIONET is a multilayer feedforward neural network with known
number of hidden nodes. Diagnosis of neck and arm pain diseases has been a
ttempted as a case study using BIONET and the results are encouraging. The
proposed model has also been tested for solving an important benchmark prob
lem known as XOR problem. In this paper due attention is given for designin
g and developing an artificial neural network model, diagnosing diseases an
d solution strategy for solving XOR problem. Major applications of BIONET i
nclude diagnosis of diseases and parity checking. Quicker convergence durin
g training is the main advantage of BIONET without loss of accuracy. (C) 20
00 Elsevier Science B.V. All rights reserved.