نمایش چکیده فارسی

Explainable Machine Learning for Cardiovascular Disease Prediction: Internal Validation, Cross-Dataset Evaluation, and a Preventive Decision-Support Framework

Parisa GhahremanShahraki1

Abstract :
Background and Objective: Cardiovascular diseases (CVDs) remain the leading cause of mortality worldwide, and the early identification of individuals at risk is a cornerstone of cardiovascular disease prevention [1]. Machine learning methods can model complex and nonlinear relationships among clinical and behavioral factors and therefore offer considerable potential for cardiovascular disease risk prediction. However, model interpretability, generalizability across different populations, and the translation of model outputs into information that can support prevention remain important challenges in the application of these methods in healthcare. This study aimed to develop and internally evaluate several machine learning models for cardiovascular disease prediction, assess model transferability to a different dataset, interpret model behavior using SHapley Additive exPlanations (SHAP), and develop a research-oriented framework for risk stratification and preventive decision support. Methods: The development dataset consisted of 70,000 records from the CVD-70K dataset. After quality control of blood pressure and anthropometric variables, 68,675 observations remained. Logistic Regression, tuned Random Forest, and XGBoost models were evaluated using a stratified 80/20 train-test split and stratified five-fold cross-validation within the training set. A transferable feature set comprising age, sex, systolic blood pressure, and cholesterol was defined, and four-feature Random Forest and XGBoost models were evaluated, without retraining, on 74,276 cleaned observations from the Yemen dataset. Because of the substantial difference in age distribution, cross-dataset evaluation was performed both on the entire Yemen dataset and on the overlapping age subgroup of 29–64 years. The full XGBoost model was interpreted using SHAP, and its outputs were used to construct operational risk strata and overlapping phenotypes of modifiable risk factors. Results: On the CVD-70K test set, XGBoost achieved an accuracy of 0.7301, an F1-score of 0.7142, and an ROC-AUC of 0.7952, while the tuned Random Forest achieved an ROC-AUC of 0.7963. Mean five-fold cross-validation ROC-AUC values for Logistic Regression, Random Forest, and XGBoost were 0.7919 ± 0.0027, 0.8019 ± 0.0032, and 0.8005 ± 0.0033, respectively. ROC-AUC values for XGBoost models using 4, 6, 8, and 10 features were 0.7902, 0.7930, 0.7954, and 0.7952, respectively. On the full Yemen dataset, ROC-AUC values for Random Forest and XGBoost were 0.7288 and 0.7272, respectively; both models achieved sensitivity above 0.92 but specificity below 0.16. In the 29–64-year subgroup, accuracy, specificity, and F1-score improved, whereas ROC-AUC decreased to 0.6641 for Random Forest and 0.6617 for XGBoost. In Yemen, Brier scores were 0.2328 and 0.2339, calibration intercepts were −0.7148 and −0.6884, and calibration slopes were 1.1604 and 1.0815 for Random Forest and XGBoost, respectively. SHAP analysis identified systolic blood pressure, cholesterol, and age as the most important contributors to model predictions. The observed prevalence of CVD increased from 25.19% in the low-risk group to 79.45% in the high-risk group. Conclusion: Machine learning models demonstrated moderate discrimination in CVD-70K, and the four-feature set retained most of the predictive performance of the full model. However, the altered performance pattern, low specificity, and calibration measures observed in the Yemen dataset demonstrated that satisfactory internal performance does not guarantee transferability to a different dataset. Age-restricted subgroup analysis improved several threshold-dependent performance metrics but did not improve ROC-AUC. Combining cross-dataset evaluation, SHAP-based interpretation, and profiling of modifiable risk factors provided a research framework for more transparent model evaluation; however, this framework is not intended as a diagnostic tool or as a source of individualized treatment recommendations.

Keywords : cardiovascular disease machine learning explainable artificial intelligence xgboost shap cross dataset evaluation transferability risk stratification


زبان مقاله : انگلیسی
محور مقاله : فناوري اطلاعات ( کلیه گرایش ها)
محل انتشار : کنفرانس ملی فناوری اطلاعات، نانوتکنولوژی، هوش مصنوعی و آینده پژوهی فناوری
آدرس وب سایت کنفرانس : itnaf.ir
کد IOI مقاله : AAHD-ABKGBK
لینک مستقیم مقاله : https://isnac.ir/AAHD-ABKGBK
نمایش چکیده فارسی : https://isnac.ir/SA/AAHD-ABKGBK

سایر مقالات ارائه دهنده
  1. هوش مصنوعی در هنر و طراحی شهری: مروری نظام‌مند بر کاربردها، چالش‌ها و افق‌های پیش‌رو
  2. مدلی پویا برای پایش لحظه‌ای آلایندگی خودروها مبتنی بر پردازش ویدئوی ترافیک و شبکه عصبی هیبریدی CNN-RNN در کلان‌شهرها
  3. A Novel Explainable Machine Learning Decision-Support Framework for Residential Building Energy Consumption Prediction and Management
  4. Predicting E-Learning Continuance Intention: Cross-Dataset Validation and Educational Intervention
  5. Explainable Machine Learning for Cardiovascular Disease Prediction: Internal Validation, Cross-Dataset Evaluation, and a Preventive Decision-Support Framework
  6. Explainable Road Traffic Crash Severity Prediction Using Machine Learning: Cross-Country Validation and Data-Driven Road-Safety Interventions
  7. Spatial Assessment of Urban Walkability Using Open Data and Multicriteria Analysis: A Case Study of Shahrekord, Iran
  8. Explainable Modeling of Urban Bus Inter-Stop Travel Time: Independent Cross-City Replication and Data-Driven Managerial Recommendations
  9. چارچوب تبیین‌پذیر شخصی‌سازی در تجارت الکترونیک مبتنی بر پیش‌بینی خرید از سبد و توصیه‌گری کالا‌به‌کالا با اعتبارسنجی خارجی
  10. Explainable Machine Learning for Bank Marketing Campaign Success Prediction and Decision-Support Recommendations
  11. سامانهٔ هوشمند هشدار زودهنگام ورشکستگی بنگاه‌ها با یادگیری ماشین تبیین‌پذیر و تحلیل عدم‌توازن داده‌ها
  12. ارزیابی انتقال‌پذیری مدل‌های یادگیری ماشین برای برآورد تقاضای بالقوهٔ شارژ خودروهای برقی: اعتبارسنجی خارجی بین‌شهری و مطالعهٔ آمادگی داده در اصفهان، ایران
  13. نقشه‌سازی اولویت مداخلات خنک‌کنندهٔ شهری با تلفیق سنجش‌ازدور و مواجههٔ فضایی: مطالعهٔ موردی اصفهان
سایر مقالات محل انتشار
  1. A Study on Design Understandability and Interaction of the therapy-specific Robots
  2. طراحی و مهندسی یک حسگر زیستی نوری چند مدی پلاسمونیک مبتنی بر تشدیدگرهای حلقه شکاف برای سنجش جامع آنتی‌ژن
  3. طراحی یک اسیلاتورکنترل شونده با ولتاژ با نویز فاز بسیار پایین و بازه وسیع فرکانسی با تکنولوژی نانولوله کربنی برای کاربردهای PLL
  4. یک چارچوب هوشمند آگاه به زمینه برای ارتباطات V2V و V2Iدر شبکه‌های خودرویی شهرهای هوشمند با استفاده از داده‌های ترافیکی شبکه سلولی
  5. An Ensemble Non-Dominated Feature Selection Method: A Case Study on Prostate Cancer
  6. تشخیص زودهنگام پوسیدگی دندان با استفاده از شبکه‌های عصبی
  7. Simulation and Performance Comparison of Torque Tube and LVDT Sensors in Displacer-Based Level Measurement Systems
  8. ارتقا سامانه ناوبری پهپاد مهاجر ۶ با استفاده از فناوری
  9. کاربرد نانوذرات مگنتیت (Fe3O4) در پزشکی (از سنتز تا عملکرد)
  10. Simulation and Modeling of Floating-Gate Non-Volatile Memory Performance: Focus on Oxide Thickness, Programming Speed, and Charge Transport Mechanisms
  11. Fraud Analytics with Big Data: A Literature Review on Machine Learning Algorithms in Bankruptcy Prediction and Financial Fraud Detection
  12. تحلیل آینده‌پژوهانه هم‌گرایی هوش مصنوعی، فناوری نانو و فناوری اطلاعات در مسیر تحول صنعتی آینده
  13. نقش تحول دیجیتال در خدمات کتابخانه‌های عمومی برای توان‌یابان جسمی و ذهنی
  14. Prediction and Optimization of Oxygen Production under Microgravity Conditions Using Machine Learning
  15. Simulation and Performance Comparison of Ultrasonic, Non-Contact Radar, and Guided Wave Radar Sensors in Level Measurement Systems