Enhancing Heart Failure Prediction Accuracy through Effective Preprocessing andPrincipal Component Analysis

Accurate prediction of heart failure is crucial for early intervention and preventative care.This study aims to improve prediction accuracy using a Heart Failure Prediction dataset of ۲۹۹ sampleswith ۱۲ distinct features and a target variable. We addressed data imbalance using the NearMissalgorithm and normalized the data to ensure uniformity. Subsequently, Principal Component Analysis(PCA) was used to distill the dataset to ۷ principal features, which, when aggregate with the originalfeatures, formed a restructured dataset. Several machine learning models were evaluated, and therandom forest algorithm emerged as the most accurate, achieving an ۸۳.۵% prediction success rate. Thisoutcome not only represents a significant improvement over previous studies [۱] but also highlights theimportance of meticulous preprocessing and feature optimization in predictive modeling.

کلیدواژه ها:

Heart Failure Prediction ، PCA ، Machine Learning ، Preprocessing ، Random Forest