Predictive Model for Chronic Obstructive Pulmonary Disease using Machine Learning Classifier

Abstract

The most severe chronic disease is Chronic Obstructive Pulmonary Disease (COPD) and it is a ceaseless respiratory malady that truly imperils human health and has high occurrence and mortality around the world. COPD is a chronic inflammatory lung disease causes obstruction airflow from lungs. It is a common preventable and treatable disease characterized by persistent respiratory symptoms and airflow limitation. The early identification and prediction of lung diseases have become a necessity in the research, as it can facilitate the subsequent clinical management of patients. The proposed prediction model detect COPD patient and healthy efficiently and accurately from standard derived features set from CT(Computed Tomography) images of COPD machine learning dataset. Our model used derived features set and trained model using machine learning classifier Stochastic Gradient Descent(SGD), Logistic regression(LR) and Multilayer Perception(MLP) applied with optimal parameter selection using standard approach and finally K-fold learning contributing to improve performance of proposed ML classifier. Overall scenario is novel approach for the prediction of COPD in these discrete features and mixed features set effects are analyzed using supervised machine learning algorithm SGD, LR and MLP. Our proposed model excellent performance and prediction accuracy on COPD dataset and with proposed LR achieved AUC of 0.9667 and MLP received AUC of 0.9695 also is outstanding performance than previously reported classifiers on same sort of dataset.