Multiple Machine Learning Intelligent Approaches for the Heart Disease Diagnosis

Abstract

Earlier prediction of any disease is a key challenge before it engulfs the other part of the body. Heart Disease (HD) is one of the most common and dangerous diseases in this decade. Its importance cannot be underestimated as it circulates blood to each and every organ of the body. Even a missed beat can be alarming, that's why many doctors and researchers are trying to find every feasible solution from every corner. Machine Learning (ML) has opened a new doorway for scientists to assist doctors in disease diagnosis. In this research work authors have implemented four ML techniques i.e., Logistic Regression, Support Vector Classifier, Naïve Bayes, and Random Forest on the HD dataset extracted from the UCI ML repository. The Min-Max normalization technique is used for data preprocessing and models are fitted on 90:10 train-test ratio. The authors evaluated the Area Under the Curve (AUC); 91.15%, 83.64%, 94.79%, & 95.93% for the LR, SVM, NB, and RF, respectively. These models are validated by the KFold cross-validation technique.

Ähnliche Arbeiten