Clinical Insight: Comparative Analysis of Deep Learning Models for Disease Prediction across Multifaceted Datasets

This research introduces an advanced model architecture for chronic disease detection, leveraging highly developed artificial intelligence algorithms. The prototype combines the Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) procedure, showcasing significant advancements in predictive capabilities. The proposed hybrid approach, integrating CNN and LSTM with forwarding selection, is adopted for disease presence prediction. Employing a tenfold cross-validation strategy and optimizing with 200 training cycles, the model demonstrates a meticulous training process. Learning rate control plays a pivotal role, and a rate of 0.01 is chosen for optimal learning. The model is applied to datasets related to cancer, heart, kidney, and liver diseases, among other diseases, with feature selection based on information gain. Seven cutting-edge classification algorithms-RF, SVM, KNN, CNN, and CNN_LSTM-are used in a thorough examination. The outcome showcases the superiority of the proposed CNN_LSTM model with notable accuracy rates-98% and 99%, respectively, for chronic kidney disease prediction.