Medical Record Analysis to Predict Heart Failure Patient Survival Using Data Mining

Abstract

Research on heart failure patient medical record analysis is a complex healthcare issue with broad implications across various sectors. The primary focus is on extracting insights from large and complex data sets, using ensemble algorithms such as Random Forest, Extreme Gradient Boosting, Extra Tree, and AdaBoost. The results showed that Random Forest performed best with an accuracy of 0.84% ​​and an AUC of 0.89 across 299 medical records sampled in this study. This indicates high effectiveness in classifying patients based on their survival potential. Data mining can significantly support evidence-based medical decision-making and improve heart failure disease management by providing deeper insights through the identification of patterns and correlations in health data. This approach enables improved patient care quality and provides methodological recommendations for future clinical practice.