A Comprehensive Study of Artificial Intelligence-based Medical Diagnosis

Necessity of near-accurate prediction algorithms in the medical diagnosis as well as prognosis fields has been ever-impending and subject to studies on diversified paradigms. The existing data on due course of medical conditions and comprehensible and conspicuous symptoms can be utilized to train an expert system to project plausible present condition(s) of the subject along with predicting the future course of the disease. This paper will focus on assessing various existing techniques of prognosis in the artificial intelligence domain, viz. Neural Networks and Fuzzy Logic, Backpropagation, Clustering, k-Nearest Neighbour, and Natural Language Processing Technique. This paper will also focus on the challenges faced by each of these techniques.