Classification of stroke patients using data mining with adaboost, decision tree and random forest models

A stroke is a fatal disease that usually occurs to the people over the age of 65. The treatment progress of the medical field is growing rapidly, especially with the technological advance, with the emergence of various medical record data sets that can be used in medical records to identify trends in these data sets using data mining. The purpose of this study was to propose a model to classify stroke survivors using data mining, by utilizing data from the kaggle sharing dataset. The models proposed in this study were AdaBoost, Decision Tree and Random Forest, evaluation results using Confusion Matrix and ROC Analysis. The results obtained were that the decision tree model was able to provide the best accuracy results compared to  the other models, which was 0.953 for Number of Folds 5 and 10. From the results of this study, the decision tree model was able to provide good classification results for stroke sufferers.