Published Paper Details:

Given its prominence as a leading cause of deaths worldwide, careful detection of this condition is still very important and current heart disease diagnosis methods are heavily dependent on experts, who however, need time to process patients and show the potential possibilities of human error. The research project is aiming to develop a deep learning-based system for heart disease prediction in order to make it efficient and accurate. To do such complex patterns extraction the system uses Convolutional Neural Networks and Long Short-Term Memory (LSTM) networks, which work on the blood pressure readings (t) together with patient's age (l) and cholesterol levels (b) as well as electrocardiogram (ECG) result. Feature selection methods are used to find acceptable information characteristics to reveal important predictive variables, improving the system understanding and the system performance at operation. Using a comprehensive dataset, the model completes training and its accuracy works to the optimal levels for it to be able to identify the patients within the appropriate risk groups. Robustness and reliability of the performance of the model has been established through the use of precision, recall, F1- score and ROC-AUC evaluation metrics. The system is designed to be working as a diagnosis tool that helps healthcare personnel to detect the heart diseases in a timely and convenient way. The proposed solution tries to minimize diagnosis delays and improve patient well-being and also assist in preventive healthcare operations through automation of predictions. A system for implementation within clinical operations could do this and change the direction of heart disease strategies by providing quick and reliable early diagnosis options to healthcare professionals.