An ensemble model for Heart disease data sets

Diagnosing Heart diseases is one of the problems that require high level of accurate analysis and prediction. Using ensemble methods in decision support systems provide an important help in analyzing this type of diseases. Different data is extracted from different research laboratories referring to the same disease. This requires further analysis and rework in detecting the best ensemble. Trying different combinations of classification techniques for every data set is not the best solution and consumes a lot of time and effort. The proposed framework seeks the best ensemble combination method suitable for diagnosing heart diseases. This ensemble is a majority vote based method and is designed for every data set belongs to the domain of heart disease. The experimental analysis is applied on two benchmark data sets extracted from two different resources. The classification accuracy results reached percentages higher than 90% accuracy. Observations reveal that the best combination for both datasets is mostly the combinations which the Naive Bayes as one of its classifiers with accuracy of 92%.