A Peer Review of a Predictive Model for Early Detection of Heart Failure Using Machine Learning
Quarshie Frederick
Statistician, Department of Statistics, Local Government Service, Greater Accra, Ghana
ABSTRACT:
Most heart failure is detected in designated health facilities using either biomarkers or the client’s history by healthcare specialists. These setbacks are attributable to the client’s willingness to visit the facility regularly for checkups to determine if any relatives have experienced heart failure or have had any predisposed conditions. This paper systematically reviewed different algorithms employed in early detection of heart failure using artificial intelligence based on attributes like age, blood pressure, family history, diabetes, biomarkers, and angina. Thus, the review provided information about existing models for early detection of heart failure, including their strengths and weaknesses. It was observed that patients’ survival stages of flow of pain before and during spark heart failure; the fluid lipid membrane filtered the stress fluid from the atriums and ventricles and accumulated in the left lower pericardial heart failure and the accurate proportionate lipid (kg/dL) in the blood required for aerobic respiration. It is observed in the review that although a significant research effort has been made on heart failure detection, a thorough literature review is still slagging and concluded that detecting heart failure in hospitals without inculcating turn-around costs, distance, disability status, proximity to healthcare, treatment level, home, orthodox diagnosis and treatment, and other treatment centers. Hence, the fusion of these attributes will help improve early detection of heart failure. Early identification of heart disease is critical because it is one of the primary causes of cardiovascular disease-related deaths worldwide. Biomarkers, biomedical care, healthcare, and disease prediction are now actively using emerging technologies like machine learning and deep learning. This research compares and analyses current prediction models with an emphasis on heart failure prediction. Metrics like accuracy, recall, and ROC curves assess supervised machine learning for prediction algorithms like K-Nearest Neighbors, Binary Logistic Classification, and Naïve Bayes. The effectiveness of these foundation classifiers was evaluated against ensemble modeling strategies including stacking, boosting, and bagging.
Published in: International Journal of Recent Advances in Multidisciplinary Topics (Volume 5, Issue 10, October 2024)
Page(s): 28-37
Date of Publication: 18/10/2024
Publisher: IJRAMT
Cite as: Quarshie Frederick, “A Peer Review of a Predictive Model for Early Detection of Heart Failure Using Machine Learning,” in International Journal of Recent Advances in Multidisciplinary Topics, vol. 5, no. 10, pp. 28-37, October 2024.