Detecting Chronic Obstructive Pulmonary Disease with Deep Learning Using Recurrent Neural Networks (RNN)

Existing system for the diagnosis of Chronic Obstructive Pulmonary Disease (COPD) mostly focus on conventional techniques such spirometry, CT scans, and chest Xrays, which have distinct limitations including reliance on patient compliance, great expenses, and restricted accessibility. These techniques may overlook chances for early management if these detect COPD only after significant symptoms have started. By means of Long Short-Term Memory (LSTM) networks, a class of Recurrent Neural Network (RNN), the proposed system analyzes temporal patient data including demographic information, clinical histories, and spirometry results. By identifying insignificant trends that traditional techniques would miss, the predictive methodology helps to early detect COPD. By means of rigorous data preprocessing, model training, and evaluation, the proposed system attained an accuracy of 92.5%, precision of 90.0%, recall of 88.0%, and an F1-score of 89.0%. The technique shows notable increases in COPD diagnosis and patient outcomes when compared to existing systems with better performance measures.