Photo Credit: Freepick

The following is a summary of “Optimizing heart disease diagnosis with advanced Machine learning models: a comparison of Predictive performance,” published in the March 2025 issue of the BMC Cardiovascular Disorders by Teja et al.

Cardiovascular disease remains the leading cause of mortality worldwide, necessitating the development of precise and efficient predictive models to improve early diagnosis and patient outcomes. Recent advancements in machine learning have demonstrated considerable potential in enhancing the accuracy and reliability of heart disease prediction. This study utilized a dataset from the UC Irvine Machine Learning Repository, comprising patient records from Cleveland, Switzerland, Hungary, Long Beach, and Statlog. A subset of seven cases, each containing 12 clinical attributes, was analyzed to assess the predictive performance of various machine learning models, including Random Forest, K-Nearest Neighbors (KNN), Logistic Regression, Naïve Bayes, Gradient Boosting, AdaBoost, XGBoost, and Bagged Trees. 

Model performance was evaluated based on accuracy, precision, recall, F1-score, and the area under the receiver operating characteristic curve (ROC-AUC). To ensure model robustness and generalizability, K-fold cross-validation was performed with K = 10 and K = 5. Among the tested models, Random Forest exhibited exceptional stability, achieving an accuracy of 94% with K = 10 and 92% with K = 5. XGBoost demonstrated slightly lower accuracy, with scores of 90% and 89% for K = 10 and K = 5, respectively. KNN showed signs of overfitting, as evidenced by a notable accuracy decline (71% for K = 10 and 72% for K = 5). Overall, XGBoost and Bagged Trees achieved the highest accuracy at 93%, followed by Random Forest and KNN at 91%. Additionally, Random Forest and Bagged Trees exhibited the highest ROC-AUC values at 95%, with XGBoost closely following at 94%. These findings underscore the effectiveness of ensemble learning techniques in predicting cardiovascular disease, highlighting their potential for improving clinical decision-making. 

Furthermore, the study suggests that future advancements incorporating hybrid models and advanced survival analysis techniques may further enhance predictive capabilities, ultimately contributing to more accurate and timely identification of individuals at risk for cardiovascular disease.

Source: bmccardiovascdisord.biomedcentral.com/articles/10.1186/s12872-025-04627-6