Heart Disease Prediction Using Machine Learning Algorithms

Abstract

Cardiovascular disease ranks among the major causes of death in the global population, and timely and proper diagnosis is one of the most important healthcare goals. Machine learning has demonstrated good potential in helping clinicians to detect some latent trends in clinical data. This paper illustrates a heart disease predictive algorithm that applies the Logistic Regression (LR) and Random Forest (RF) algorithms. It uses a pipeline of data preprocessing, feature encoding, model training, cross-validation, and performance evaluation that is structured. Logistic Regression is interpretable and gives probabilistic results, whereas the random forest is more accurate by means of ensemble learning and non-linear modeling. The experimental analysis proves that the Random Forest tends to gain higher classification results, whereas the Logistic Regression can be still considered a reliable and explainable 9eline. The accuracy, precision, recall, F1-score, and ROC-AUC are also highlighted in the study as assessment measures that would be clinically relevant.