EARLY HEART ATTACK PREDICTION BY USING MACHINE LEARNING

Abstract

Machine learning has been applied in different fields of life it is tremendously excelling in every field of life, exploring the secrets of life, enhancing the way of living, predicting new trends and trajectories, and paving the way for a new and advanced global home. ML is also a main contributor to the digital healthcare system. By using ML models healthcare professionals have been diagnosing different health issues and by using this leverage, we have been using ML models to predict the early threat of heart attack to support the healthcare system as well as human well-being. Heart Attack is one of the growing death concerns for Homo sapiens. Many people die every day owing to cardiac attacks which lead to devastating consequences. It often goes undetected until a critical event occurs, highlighting a pressing need for early and accurate risk assessment. This study aims to recognize entities at high jeopardy of cardiac events with an exceptional degree of accuracy. Prominently, the system is planned to alert users—encouraging them to consult with senior doctors promptly. For this study, we have used a publicly available secondary “Heart Attack Prediction” data set. For the training of the model of ML, we have torn apart the data set into a training set and a testing set with ratios of 90% and 10% respectively. For parameter fine-tuning, we used Python’s Optuna library for the best ML model and its parameters. We found that the Light GBM classifier is best for binary classification and it shows promising benchmarks on test data.