A Novel Data Analytics and Machine Learning Model Towards Prediction and Classification of Chronic Obstructive Pulmonary Disease

Abstract

In today’s world, people are facing numerous infectious diseases because of varying climate and their changing lifestyles. So, early prediction of disease has become an essential task. Yet, it becomes too difficult for doctors to precisely determine the disease correctly for proper diagnosis. For solving this issue, data analytics and data mining play a significant role in disease prediction. Medical science has large amounts of data and the volume of data keeps increasing annually. Owing to the rise in data growth in scientific and healthcare sectors, reliable analysis methods of medical information is considered vital. Data analytics based on data mining discover a secret pattern knowledge over vast numbers of patient data using disease data. Chronic obstructive pulmonary disease (COPD) was identified to be a chief global public health issue due to its associated impairments and increased death rate. COPD has become a common chronic disease with certain major additional pulmonary consequences that might add to the severity of the affected individuals. Moreover, it was identified as a significant public health concern and a leading source of death globally. Moreover, COPD is normally not identified and treated in the early stage. A risk model needs to be developed for predicting the progression of COPD. The research work in this chapter employs Data Analytics and Machine Learning (ML) algorithms to analyze the COPD data. We equate all of the ML algorithms' predictive power to COPD disease. Based on the important variables, we forecast the test data and measure the prediction performance. The algorithms Random Forest and K-nearest neighbor offered best results in the analysis of COPD data, when compared to its peers.