Estimating the Chances of Getting Heart Disease using Machine Learning Algorithms

Abstract

One of the deadliest illnesses that cause death is heart disease. Worldwide, almost 17 million people died each year because of various heart diseases. To aid in the early diagnosis of heart illness, improved diagnosis, high-risk individuals, and enhanced decision-making for extra treatment and prevention, a prediction model can be proposed. Many academics have looked at the heart disease risk variables and suggested certain machine learning algorithms. However, these models need to be enhanced in order to produce findings that are extremely precise due to the large dimensionality of the data. This study intends to develop a novel framework for accurate heart disease diagnosis. The proposed model can generate precise data for the training model by applying effective approaches for data collection, pre-processing, and transformation. The proposed model employs a combined dataset from the universities of Switzerland, Hungarian, Cleveland, Long Beach VA. This model employs Relief methods for feature selection. Ensemble learning is used to generate novel hybrid classifiers. The outcomes demonstrated that hybrid classifiers performed better than current models that displayed an accuracy of above 95%. These results suggests that the model with relief feature selection and hybrid classifiers may be a more effective approach for predicting heart diseases.