Deep Learning Based Healthcare Method for Effective Heart Disease Prediction
Loveleen Kumar,C Anitha,Venka Namdev Ghodke,N Nithya,Vinayak A Drave,Azmath Farhana
DOI: https://doi.org/10.4108/eetpht.9.4283
2023-10-31
EAI Endorsed Transactions on Pervasive Health and Technology
Abstract:In many parts of the world, heart disease is the leading cause of mortality diagnosis is critical Towards Efficient Medical Care and prevention of heart attacks and other cardiac events. Deep learning algorithms have shown promise in accurately predicting heart disease based on medical data, including electrocardiograms (ECGs) and other health metrics. With this abstract, Specifically, we advocate for deep learning algorithm in accordance with CNNs for Deep Learning effective heart disease prediction. The proposed method uses a combination of ECG signals, demographic data, and clinical measurements Identifying risk factors for cardiovascular disease in patients. The proposed CNN-based model includes several layers, such as convolutional ones, pooling ones, and fully connected ones. The model takes input in the form of ECG signals, along with demographic data and clinical measurements, and uses convolutional layers to get features out of raw data. To lessen the effect of this, pooling layers are dimensionality of the extracted features, while layers that are already completely linked to estimate the risk of cardiovascular disease based on the extracted features. Training and evaluating the suggested model, We consulted a broad pool of ECG signals together with patient clinical data, both with and without heart disease. Training and test sets were created from the dataset testing arrays, and the prototype was trained using backpropagation and stochastic gradient descent. The model was evaluated using standard quantitative indicators such the F1 score, recall rate, and accuracy rate. The outcomes of experiments demonstrate the suggested CNN-based model achieves high accuracy in predicting heart disease, with an overall accuracy of over 90%. The model also outperforms several alternatives to classical techniques for heart disease prediction, including the more conventional forms of AI algorithms different forms of deep learning models. In conclusion, the proposed deep learning algorithm based on CNNs shows great potential for effective heart disease prediction. The model can be integrated into healthcare systems to provide accurate and timely diagnosis and treatment for patients with heart disease. Further research can be done to optimize the model's performance and test its effectiveness on different patient populations.