Two Majority Voting Classifiers Applied to Heart Disease Prediction

Two novel methods for heart disease prediction, which use the kurtosis of features and Maxwell–Boltzmann distribution, are presented. A Majority Voting approach is applied, two base classifiers derived through statistical weight calculation. First, exploitation attribute Kolmogorov–Smirnov test (KS test) result done by plugging categorizer into a Bagging Classifier. Second, fitting Maxwell random variables to components summating KS statistics used assignment. We have compared state-of-the-art proposed reported results. According findings, our Gaussian distribution kurtosis-based Classifier (GKMVB) Distribution-based (MKMVB) outperform SVM, ANN, Naive Bayes algorithms. In this context, also indicates, especially when we consider that hack intuitive, routine promising. Following state-of-the-art, experiments were conducted on well-known datasets Heart Disease Prediction, namely Statlog, Spectf. comparison Optimized Precision made prove effectiveness methods: newly attained 85.6 81.0 Statlog Spectf, respectively (while state 83.5 71.6, respectively). claim family still open new developments appropriate This obvious, its simple structure fused with Ensemble Methods’ generalization ability success.