Artificial neural network (ANN) has become a strong tool to solve the
problems of considerable complexity i.e. problems that do not have an
algorithmic solution or for which an algorithmic solution is too complex
to be found. An artificial neural network (ANN) is an
information-processing paradigm and is inspired, by the way biological
nervous systems works, such as the brain process information. ANN are
abstracted from the biological brain & therefore problems that are too
complex for conventional technologies but people are good at solving such
as pattern recognition, forecasting (which requires the recognition of
trends in data) are well suited to ANN. ANN’s most important advantage in
solving such problem over brain is that their capability is not affected
by factors such as fatigue, working conditions, emotional state, and
compensation.

Diagnosis prediction is one such problem, where human are better in
solving but computers are not. Expert system based medical diagnosis is
highly an active research field now a days. ANN is applied and found
useful in predicting the disease to a considerable limit from neural
network knowledge base framed with the help of past patients data. Perfect
diagnosis, effective treatment and high quality of medical care by a
physician is the basic aim. This application is significant in helping the
physician where large amount of data has to be considered. Because every
human has its own perception depending upon the information stored in his
brain, thus where ever large amount of data has to be considered many a
times different medical practitioners has different opinion.

The medical expert system makes use of neural network as their knowledge
base for predicting disease. Patient data and diagnosis is processed into
a
usable format and disease vectors (neurons) are formed. Neural networks
are trained with disease vector consisting of patient’s features
(patient's data i.e. symptoms and complaints) as their dimensions. Several
cases of different disease are taken into consideration. Each disease
vector is a prototype (neuron) for the neural network. System is made to
learn these prototypes and test vectors are processed to predict the
disease patient is suffering from.

More vectors can be added to enhance the neural network knowledge base.
System can be updated by adding each new case in order to have a wealth of
information from the past patient data.