A Collaborative Empirical Analysis on Machine Learning Based Disease Prediction in Health Care System

Medical treatment processes around the world are expected to revolutionize with the help of AI-aided healthcare services. AI can replicate human cognition and is capable of learning, reasoning, making decisions, and acting. The adaptation of AI can radically reshape the entire healthcare businesses. This paper proposes a comparative analysis of four classification algorithms. The selected algorithms are k-Nearest Neighbour, Naive Bayes, Decision Tree, and Random Forest which predict some commonly identified ailments. These Supervised Machine Learning classifiers are used on a disease prediction data-set to predict 41 prevalent diseases based on any 5 conspicuous symptoms from the data set’s 132 common symptoms. After the study we conclude, Random Forest had the highest accuracy of 99.5%, followed by Decision Tress at 95.8%, then K Nearest Neighbor at 93.4%, and lastly Naive Bayes at 87.7%. Our experiment achieved higher accuracy than previous investigation dealing with identical classification problem, in the health care arena. A web application has also been employed which is providing the highest accuracy while predicting the diseases in a user-friendly manner.