Heart Failure Prediction Improvement Using Fuzzy Logic Pre-processing and Random Forest Classifier

Heart failure (HF) is a critical condition that necessitates precise prediction for effective intervention. Traditional machine learning approaches have shown promise in predicting heart failure, yet their performance can be hindered by noisy and imprecise data. This study explores the integration of fuzzy logic pre-processing with a Random Forest classifier to enhance the prediction accuracy of heart failure outcomes. The objective of this work is to determine how well a Random Forest classifier predicts heart failure when pre-processed with fuzzy logic. Could fuzzy logic pre-processing improve machine learning models' ability to predict medical prediction. This is the research subject that is being tackled. In order to convert raw input characteristics into more informative fuzzy features, fuzzy logic pre-processing is applied in the methodology. The classification accuracy of the Random Forest model is then compared with and without this pre-processing. Crucial results indicate that adding fuzzy logic pre-processing led to a marginal increase in accuracy, from 76.67% to 77.78%. The primary finding indicates that pre-processing with fuzzy logic has a beneficial effect on model performance, indicating that it may improve medical predictions. Fuzzy logic, with its capability to handle uncertainty and imprecision, can preprocess the data, making it more suitable for subsequent classification tasks. Furthermore, we provided key performance metrics such as accuracy, precision, recall, and F1-score for the models tested (Gaussian Kernel and Bayesian Inference before and after balancing). We present a case study demonstrating the effectiveness of this hybrid approach.