SDCANet: Enhancing Symptoms-Driven Disease Prediction with CNN-Attention Networks

Deep learning algorithms have revolutionized healthcare by improving patient outcomes, enhancing diagnostic accuracy, and advancing medical knowledge. In this paper, we propose an approach for symptom-based disease prediction based on understanding the intricate connections between symptoms and diseases by accurately representing symptom sets, considering the varying importance of individual symptoms. This framework enables precise and reliable disease prediction, transforming healthcare diagnosis and improving patient care. By incorporating advanced techniques such as a one-dimensional convolutional neural network (1DCNN) and attention mechanisms, our model captures the unique characteristics of each patient, facilitating personalized and accurate predictions. Our model outperforms baseline methods through comprehensive evaluation, demonstrating its effectiveness in disease prediction.