Logistic Regression's Effectiveness in Feature Selection with Information Gain in Predicting Heart Failure Patients

Abstract

Listen

Translate article

Heart failure is a chronic illness that obstructs blood flow, which is necessary for the body to circulate oxygen. Patients with heart failure have a poor chance of survival, as evidenced by the high death rate. The hospital's infrastructure and medical facilities determine the degree of patient safety, and the patients' medical records play a significant role in ensuring that they receive the right care. As a result, a system that uses specific data to forecast the safety of heart failure patients is required. Machine learning, a computer-based approach, is one way to get around this. The logistic regression algorithm has been used to generate predictions in earlier studies. The approach for feature selection from the dataset that is suggested in this study is information gain. You can filter features that are significant to the dataset in this way. In addition, selection can enhance machine learning efficacy by decreasing the dimensions of the data. Five features—time, serum creatinine, ejection fraction, age, and serum sodium—are the outcome of information gain. After that, predictions were made using logistic regression, and a data sharing ratio of 70% training data and 30% test data resulted in an accuracy of 0.8556. This demonstrates how feature selection with Information Gain can improve the accuracy of the logistic regression model and is a very effective method.