Empirical Comparison of Decision Tree, Logistic Regression and Support Vector Machine
Three classification algorithms were considered (Decision Tree, Logistic Regression and Support Vector Machine) in the prediction of survivors in the Titanic disaster (focus is not on dataset). The performances of three models created were evaluated. The purpose of this research was to point out features of these learning algorithms that most influenced their individual predictive capabilities and to propose hypothetical combination of features of these three algorithms to form a hybrid algorithm with better predictive capabilities compared to the individual algorithms studied. A future research can make use of the result of this study to implement a new Learning Algorithm. Open in Studio to see the Experiment.