Hybrid K-Means and Support Vector Machine to Predict Heart Failure

Prevalence of Heart Failure (HF) is increasing globally every day in addition to many undiagnosed cases. Affecting at least 26 million people worldwide, this chronic condition can’t be cured while it can be prevented. Over the years, researchers have come up with machine learning models to predict HF in an attempt to help diagnosticians to be better equipped. This paper proposes a prediction by applying K-Means clustering algorithm on the UCI Heart Failure Clinical Records data, thereby classifying the entire data into six clusters. Further, Support Vector Machine (SVM) is used on this cluster data for prediction of HF in patients based on risk factors which was achieved with an accuracy of 93.33%. Tuning the hyperparameters using grid search enhanced the model predictions.