A Pragmatic Comparison of Supervised Machine Learning Classifiers for Disease Diagnosis

This study focuses on comparing the different supervised machine learning classifiers such as Logistic Regression, Naïve Bayes, Support Vector Machine, K-Nearest Neighbour, Decision Tree, Random Forest, AdaBoost and Multilayer Perceptron used for diagnosing and predicting the diabetes disease from the hepatic and lipid profile panel and choosing the most suitable method based on the accuracy of these algorithms. The research study essentially puts forward a novel approach to predict one disease from the markers of other related diseases. The dataset used in all the experiments mentioned in the paper has been collected from a medical center situated in Kashmir for a period of six months. Jupyter Notebook has been used as a data analytics tool and Python 3.7 as a programming language to perform all the experiments. The data for training and testing has been split into different ratios of 80:20 and 70:30. The observed results determine that the Random Forest algorithm attained the best accuracy of 82.72% with only nine predictive attributes.