Heart failure, a severe clinical condition often resulting from underlying cardiovascular diseases (CVDs), is a leading cause of global mortality. Accurate and early prediction of patient outcomes is crucial for implementing timely medical interventions and improving survival rates. This study aims to enhance the classification accuracy of heart failure patients by leveraging optimized machine learning algorithms. The dataset used comprises 12 critical clinical features, including age, blood pressure, ejection fraction, and serum creatinine, which are instrumental in assessing a patient’s risk level. To achieve optimal prediction performance, two distinct models are employed: a traditional Decision Tree Classifier and a deep learning-based Long Short-Term Memory (LSTM) network. To further refine the LSTM model, Particle Swarm Optimization (PSO) is integrated, which effectively tunes hyperparameters and enhances the model’s convergence and generalization ability. The combination of PSO and LSTM demonstrates a robust approach to learning complex patterns in the dataset and improving prediction accuracy. The models are evaluated using standard performance metrics such as accuracy, precision, recall, and F1-score. Experimental results indicate that the optimized LSTM-PSO model outperforms traditional classifiers in identifying patients at high risk of mortality due to heart failure. This predictive system can aid healthcare professionals in early diagnosis, personalized treatment planning, and ultimately, reducing the burden of cardiovascular diseases through data-driven decision support. 

  Keywords: Heart Failure Prediction, Cardiovascular Disease, Decision Tree, LSTM, Particle Swarm Optimization (PSO), Machine Learning, Deep Learning, Mortality Classification, Healthcare Analytics, Early Diagnosis.