A Review of Machine Learning’s Role in Medical Diagnosis, Revised and Updated

Abstract

This study surveys the evolution of machine learning-based intelligent data analysis in healthcare, including its origins, current
state of the art, and potential future directions. The goal of the article is not to provide a broad overview, but rather to highlight
some subfields and future prospects that, in my opinion, will be crucial for medical diagnosis using machine learning. My focus
in the historical review is on decision trees, neural networks, and the naïve Bayesian classifier. I provide a comparison of many
medical diagnostic tasks performed by state-of-the-art systems that represent each area of machine learning. Two examples
show the patterns that will emerge in the future. The first one talks about a new approach to dealing with classifier reliability that
might be useful for medical intelligent data analysis. While complementary medicine is not (now) recognized by conventional
medicine, it has the potential to become an integral part of medical diagnosis and treatment in the future. The second article
explains a method to use machine learning to validate some inexplicable events in this field.