Heart Disease Diagnosis using a Machine Learning Algorithm

Machine Learning turns the extensive collection of raw healthcare data into information that can help to make informed decision and prediction. This research named fine-tune prediction (FTP) model aims to identify significant features and incorporate hybrid classifier to improve accuracy. For this, the Cleveland heart samples taken from the UCI repository. Experiment results show that the proposed FTP model achieves 93.49% accuracy. Without FTP, the RF, and LM achieve only 88.20% and 63.60% accuracy. We also present the hyper-parameter tuning of the classifier with significant features in the result section.