Sci Rep. 2026 Apr 27. doi: 10.1038/s41598-026-47114-4. Online ahead of print.
ABSTRACT
Cardiovascular Disease (CVD) risk assessment involves evaluating various clinical and lifestyle factors to estimate an individual's likelihood of developing heart-related conditions. Accurate risk prediction helps in early intervention and preventive care. Min-max scaling is applied during pre-processing to ensure all input features are normalized to a standard scale, typically within a given range. Feature selection plays a crucial role in improving the performance of machine learning models, especially when dealing with high-dimensional datasets like those used in CVD risk assessment. This research proposes a hybrid feature selection framework combining Artificial Bee Colony (ABC) optimization with Decision Tree classifiers. The ABC algorithm identifies the optimal subset of features by simulating the food-foraging behavior of bees, efficiently exploring the feature space, and selecting the most relevant features for classification. ABC's exploration and exploitation capabilities help avoid local optima and ensure a robust feature selection process. Once the optimal features are selected, we employ several Decision Tree-based classifiers, including the C4.5 algorithm and Random Forest, to predict the risk of CVD. The C4.5 algorithm is widely used for classification tasks because it generates human-readable decision trees. At the same time, Random Forest, an ensemble approach, combines the predictions of many decision trees to improve accuracy and robustness against overfitting. The framework's performance is measured using criteria such as accuracy (94.2%), precision (93.5%), recall (94.8%), F1-score (94.1%), and specificity (95.2%) to determine the usefulness of CVD risk prediction. These measures provide a fair assessment of the model's capacity to identify both at-risk and healthy people properly.
PMID:42045379 | DOI:10.1038/s41598-026-47114-4