Development of novel predictive model for chronic disease using machine learning

Abstract

Prediction of chronic disease plays a key role in the healthcare field. Diagnosing the disease at an early stage is important. Due to the considerable technological progress, enormous data volumes are produced in computer science. The progress of the clinical information networks produces numerous medical databases. The extraction of knowledge and management of large volumes of heterogeneous data has become a key research field, known as data mining. The precise review of medical data benefits early identification of a disease, patient care as well as community services from large data development in biomedical and healthcare communities. The extraction and management of knowledge from large volumes of heterogeneous information have now become a key research area. The accurate processing of health data benefits early identification of disease, patient care and community services by means of large data increases in the biomedical and healthcare communities. However, if the consistency of medical data is insufficient, analytical precision is diminished. The consistency of the perception and diagnosis of chronic disease is guaranteed for the fields of biomedical pattern recognition and master learning. The objective of the decision-making approach is also promoted. Machine learning provides a powerful solution for the study of bio-medical information that is both high-dimensional and multi-modal. Prediction of chronic diseases plays a key role in computer science. It is essential to predict the chronic disease at an early stage. The proposed model analyses on the chronic obstructive pulmonary disease dataset. Our work introduces chronic obstructive pulmonary disease prediction system with supervised machine learning techniques like Stochastic Gradient Descent, Logistic Regression, Multiplayer Perceptron, Random Forest and XG boost. We then analyse classification methods with many parameters for chronic disease prediction including AUC, ROC, precision, sensitivity and accuracy.