Published Paper Details:

Unfortunately, there is currently no cure for coronary heart disease (CHD), a deadly cardiac condition. Identifying coronary artery disease accurately and quickly is crucial for effective patient therapy. Treatments can be initiated sooner and patient outcomes can be improved with early detection. Using an optimised LightGBM classifier, the "HY_OptGBM" model forecasts CHD. When it comes to predictive modelling, the gradient boosting framework LightGBM is both efficient and accurate. Optimisation of the LightGBM classifier is achieved by adjustments to the hyperparameters and loss function. The accuracy and efficiency of model training are both enhanced by this optimisation strategy. Data on coronary heart disease from the Framingham Heart Institute is used to evaluate the model's performance. The algorithm reliably predicts CHD using this data, which paves the way for early detection and, maybe, reduced treatment expenses. Additionally, it presents a Voting Classifier (RF + AdaBoost) that can identify Coronary Heart Disease with a 99% accuracy rate. This ensemble model using Random Forest and AdaBoost does a good job of differentiating CHD patterns. User registration and sign-in are made easier with the use of an intuitive Flask framework that integrates with SQLite, allowing for easier usability testing. This streamlined interface may make the use of machine learning technologies easier for the stakeholders involved in CHD identification.