Predicting COPD Severity Using Machine Learning and GOLD Criteria

Breathing problems that get worse over time are a symptom of chronic obstructive pulmonary disease, or COPD. It is a catch-all phrase for a number of conditions including asthma, bronchitis, and emphysema. It is defined by a progressive constriction of the airways and an early aging of the lungs. Experts have attested to the fact that early identification helps lessen the symptoms of COPD. Thus, using GOLD criteria, patients' COPD severity can be predicted using contemporary methods like machine learning algorithms. The proposed prediction models predict severity of COPD of the patient on a scale of 1 to 4 as prescribed by the GOLD system. Our model was trained to utilize a derived feature set and machine learning classifiers, such as SVC, Random Forest Classifier, XGBoost classifier, Naive Bayes, and Hard Voting Ensemble, with the best possible parameter choice to achieve a commendable accuracy of 97.6%. The global COPD epidemic will be lessened with the use of this method and medical professionals' experience.