Ensemble-Based Prediction of Myocardial Ischemia Complications

Myocardial ischemia, a condition where the heart muscle doesn’t receive enough blood and oxygen, is a serious concern worldwide. It can lead to life-threatening complications such as arrhythmias, cardiac failures, and secondary infections. These complications not only increase the risk of death but also result in increased morbidity and increased healthcare costs. This highlights the need for a more proactive approach in identifying high potential risk patients developing myocardial ischemia complications. To tackle this problem, a research experiment has been initiated which enables ensemble learning techniques to predict myocardial ischemia complications. This paper aims to use patient demographics, medical history, laboratory test results, and other relevant information to develop a model that can accurately predict these complications. By doing so, healthcare providers can identify high-risk patients early, allowing them to make informed treatment decisions and provide timely intervention, ultimately leading to improved patient outcomes and reduced healthcare costs. This work proposes a combination of Multi-layer Perceptron and Random Forest to classify Myocardial Ischemia complications. The Neural Network model captures complex relationships in the data, while the Random Forest model captures simple, interpretable relationships. The results of both models are combined into an ensemble model to improve performance and reduce the risk of overfitting. The total average of the accuracies for the prediction of 5 complications came out to be 90.93. The final model is deployed in a real-world setting to classify Myocardial Infarction complications. This will allow healthcare providers to take a proactive approach to care for patients, ultimately leading to improved patient outcomes and reduced healthcare costs.