Comparative Study of Supervised Machine Learning Algorithms for Healthcare Dataset Using Orange

Machine learning algorithms are special computer programs that improve their efficiency through experiences. It is the combination of statistics and computer algorithms that is used for identifying the hidden pattern and forecasting. In this paper, seven supervised machine learning algorithms are compared with three datasets of heart disease patients, diabetes patients, and depression patients. These seven algorithms are decision tree, logistic regression, K-nearest neighbor, support vector machine, Naïve Bayes, random forest, and Adaboost are implemented on Orange. Orange is an open-source graphical user interface platform for the implementation of machine learning algorithms. From the study, it is found that logistic regression and Naïve Bayes algorithm shown better results as compared to other algorithms of average accuracy of 81.23% and 79.65%. Support vector machine does not fit with these types of classification and performed accuracy of 56.34%. As an inference, SVM should not be implemented on such classification problems.