There is interest to develop computers that help doctors in diagnosis since the time of their invention. Much of the research in this is confined to some universities in US. Typically graduate students in Medical Informatics work on algorithms to generate a diagnosis which is tested by doctors. Such type of research showed Bayesian probability is promising in its application. However, the iterations involved in such calculations are prohibitively costly since a super computer is needed for one diagnosis and it takes about 50years.
However, approximation (variational) methods are developed to solve the problem using a PC. However there are several problems involved in making accurate diagnosis. Typically computational methods for medical diagnosis use a database of disease feature links which are quantified (0.09 to 0.99) based on their pathophysiology. The import of a feature given…
