Main Article Content
Stroke is an important health outcome in terms of morbidity, disability, mortality, and social and economic costs [2-3]. Method, in this paper, we build a smart prediction model which predicts whether a patient is at high risk or low risk of heart stroke with an early intervention by classifying the patient’s records into one of the binary classes. We are using K-means Clustering technique and Classification techniques and to enhance the performance in predicting accurate values, we are integrating all 3 classification algorithms – Naive Bayes NB, Decision Tree DT and Artificial Neural ANN Network, with Sequential Row initial centroid selection methods of K-means clustering algorithm.Comparison analysis of each model is determined by calculating Sensitivity, Specificity and Accuracy using Confusion Matrix of each one. We also plotted ROC and AUC score as final assessment, in choosing the best prediction model. Conclusions, we developed smart and highly accurate predicting heart stroke system. The most effective model is Sequential Row Method integrated with Artificial Neural Network Classifier with accuracy score of 96% and area under the ROC curve (AUC) score of 1.