Key result
After hyperparameter optimization, the XGBoost model achieved a higher accuracy of 91.85% compared to 87.78% for Logistic Regression in the early diagnosis of coronary artery heart disease.
Why the study?
Coronary artery heart disease fatality can be minimized by early diagnosis, but symptoms are often ignored by patients and clinicians.
Do optimized supervised machine learning models (Logistic Regression and XGBoost) improve the accuracy of early diagnosis of coronary artery heart disease compared to non-optimized models?
Do optimized supervised machine learning models (Logistic Regression and XGBoost) improve the accuracy of early diagnosis of coronary artery heart disease compared to non-optimized models?
Absolute Event Rate: 91.85% vs 87.78%
Hyperparameter optimization significantly improves the accuracy of machine learning models like XGBoost and Logistic Regression for detecting coronary artery disease.
May aid CAHD detection research; leaves open clinical adoption pending prospective validation.
The mortality rate of cardiovascular disease is increasing at an alarming rate and affects more than 31% of the world's population. The increment in obesity, strain, high blood pressure, concern, high body mass index, depression, high glucose level, and high cholesterol are the key factors that increase the risk of cardiovascular disease. Coronary heart and cerebrovascular diseases cause 80% of cardiovascular diseases with a higher mortality rate, out of which Coronary Artery Heart Disease (CAHD) causes 25% of the fatality. CAHD fatality can be minimized by early diagnosis. However, due to common symptoms of the non‐fatal disease, CAHD is often ignored by the patients and the health care team. In this paper, the authors comprehensively analysed and compared the two supervised Machine Learning models, i.e., Logistic Regression (LR) and XGBoost, for the early diagnosis of CAHD. The models are applied to a benchmark dataset (Statlog heart disease dataset). The model's parameters are also optimized by performing the Random SearchCV hyperparameter tuning. The comprehensive study is performed on both non‐optimized and optimized models as well as compared to the other well‐known existing models. After parametric optimization, both the models exhibit higher accuracy, i.e., LR 87.78% and XGBoost 91.85%, compared to the non‐optimized techniques, i.e., LR = 83.33% and XGBoost = 77.77%.
No takes yet. Share an insight, caveat, or question.
Dhanka et al. (2023) studied Coronary Artery Heart Disease (CAHD). XGBoost model (optimized) vs. Logistic Regression model (optimized) was evaluated on Model accuracy. After hyperparameter optimization, the XGBoost model achieved a higher accuracy of 91.85% compared to 87.78% for Logistic Regression in the early diagnosis of coronary artery heart disease.
Synapse has enriched one closely related paper. Consider it for comparative context: