Heart failure stands as a major worldwide health problem that requires immediate and precise diagnosis to enhance patient treatment results. Researchers developed a machine learning system to predict heart failure based on standard clinical data collected from patients. The researchers analyzed 918 patient records, which included information about patient demographics, physiological measurements, and electrocardiographic data. The analysis utilized nine variables, including age and sex, chest pain type, resting blood pressure, cholesterol level, fasting blood sugar, resting ECG results, maximum heart rate, exercise-induced angina, ST depression (Oldpeak), and ST segment slope.

The Random Forest classifier received preprocessed data through imputation, normalization, and one-hot encoding before achieving 88.59% accuracy, with precision (positive class) at 91.00%, recall at 90.00%, and F1 score at 90.00%. The model demonstrated balanced performance, with 67 true negatives, 95 true positives, 10 false positives, and 12 false negatives, as indicated by the confusion matrix.

The proposed model demonstrates that artificial intelligence can perform automatic heart failure diagnosis by providing a reliable risk assessment for early detection. The model achieves better performance through feature optimization and ensemble learning methods, while its ability to process various types of clinical data makes it suitable for real-world applications. The proposed method enables researchers to develop AI-based clinical decision support systems that assist doctors in making prompt medical decisions through early disease detection.

Future research should aim to enhance model interpretability while validating the model across different population groups and implementing explainable AI techniques to foster clinical trust and transparency, ultimately leading to improved translational outcomes.