Chronic Obstructive Pulmonary Disease Diagnosis with Bagging Ensemble Learning and ANN Classifiers

Abstract

Chronic Obstructive Pulmonary Disease (COPD) is a persistent respiratory disease that poses a significant threat to global human health with elevated incidence and mortality rates. Timely recognition and diagnosis of COPD play a pivotal role in efficiently managing and treating the condition. The incorporation of deep learning technologies into healthcare has significant potential to enhance diagnostics and treatment outcomes. This study proposes an innovative deep-learning approach along with an ensemble technique to address the imperative need for an effective predictive model in COPD disease classification, particularly in situations with limited available data. This was achieved by leveraging the ensemble bagging technique and incorporating ANN as a classifier within this framework. Training and evaluation of the proposed ensemble ANN model were performed on a dataset comprising a variety of attributes, including demographic information, medical history, diagnostic measurements, and pollution exposures. Data were collected from people aged 18 to 60 originating from Pakistan, encompassing patients, attendants, hospital staff, faculty, and students. The effectiveness of the model in classifying COPD was measured using F1 score, recall, precision, and accuracy. The evaluation of the model produced notable results, as it achieved a 90% F1 score, 96% recall, 84% precision, and 89% accuracy in identifying the presence of COPD in individuals. Furthermore, this study carried out a comparative analysis between a standalone ANN model and the proposed ensemble ANN model which revealed that the proposed Ensemble ANN model outperforms existing methods, particularly in scenarios with limited sample size. This research provides substantial contributions to healthcare technology, as it presents an efficient tool for COPD prediction, facilitates early intervention, and significantly increases the overall standard of patient care.