Ensemble Methods Based Feature Selection For Cardiovascular Diseases

Abstract

Abstract: In the healthcare sector, developing the feature selection model to predict cardiovascular diseases (CVD) is a challenging task. The potential risk factors and predictors and carefully considering feature selection techniques are necessary to create dependable prediction models. This paper explores the significance of feature selection approaches in enhancing the CVD prediction models. Although traditional machine learning models with individual methods have been used for feature selection, but their performance is not so much appropriate for the various data sets, specifically healthcare datasets. Thus, we have taken ensemble machine learning methods to solve such issues for the healthcare dataset. We have considered the ensemble machine learning model based on the combination of two methods, such as the combination of support vector machine (SVM) and random forest (RF) as (SVM+RF), decision tree (DT) and random forest (RF) as (DT+RF). We demonstrated our model using the CVD dataset, and it performed better than traditional methods. Our model is more effective in selecting the correct features from the data set and predicting the CVD.