<![CDATA[Coronary artery disease (CAD) continues to be a major global health issue, demanding more effective diagnostic techniques. This study introduces a detailed framework for CAD detection that integrates data preprocessing, feature engineering, and model optimization to enhance diagnostic accuracy. Our methodology encompasses comprehensive data cleansing to eliminate inconsistencies, transformations for better feature representation, feature reduction to highlight relevant variables, data augmentation for balanced class distribution, and optimization strategies to boost model performance. We employed a random forest classifier, trained via 5-fold cross-validation, to develop a robust model. The efficacy of this model was tested through two key experiments: firstly, by comparing its performance on preprocessed versus raw data, and secondly, against previous studies. Results demonstrate that our model significantly surpasses the one trained on raw data, achieving an accuracy of 93.00% compared to 86.16%. Moreover, when compared with existing research, our random forest model excels with an accuracy of 93.00%, a F1 Score of 93.00%, and a recall of 94.00%. Despite the superior precision of the Hybrid PSO-EmNN model found in other research, our results are promising. They underscore the potential of advanced feature engineering to further refine the effectiveness of CAD detection models. The study concludes that meticulous data preprocessing and model optimization are crucial for enhancing CAD diagnostics. Future research should focus on incorporating more sophisticated feature engineering techniques and expanding the dataset to improve the model’s precision and overall diagnostic capabilities.]]>
Copyrights © 2024
