Heart Failure Prediction Using Feature Selection Methods and Machine Learning Algorithms

When the heart is unable to pump blood as effectively as it ought to, heart failure occurs. Breathing difficulty is frequently brought on by this because blood frequently backs up and fluid can accumulate in the lungs. It is one of the major killers in the modern world at this moment. Heart failure prediction is a substantial challenge for clinical data analysis. Heart failure is a complex medical condition that might be difficult to predict using traditional risk assessment tools. Given the significant morbidity and mortality associated with heart failure, more precise methods of disease prediction are required. The first problem that has to be overcome is that one should be alerted far enough in advance to take precautions. As a result, we propose a comparative machine learning framework based on eight distinct classification algorithms, including K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Decision Trees (DT), Random Forest (RF), Logistic Regression (LR), Gaussian Naive Bayes (GNB), Multinomial Naive Bayes (MNB), and Bernoulli Naive Bayes (BNB). This study also emphasizes data visualization to spot outliers, trends, and patterns in a sizable data collection. For ML, Scikit-learn and Python are utilized. Model training is carried out using Python and Scikit-learn. Among all models, SVM achieved a test recall of 0.92 using 10 -fold cross-validation and a test accuracy of $\text{8 8. 5 9 \%}$ using the holdout method.