Prognosis of Chronic Diseases Using Different Machine Learning Models

Abstract

Chronic diseases are the diseases that persist for a long time these are classified as the conditions persist for one year or more and require regular monitoring and ongoing medical attention on a day-to-day basis. Prediction of such chronic disease in human beings can be very useful for preventing these situations and know the complication in advance; the prediction model would give information such as the disease that patients might prone to, the more likelihood of the disease that might encounter in near future. The basic idea of this prediction model is to analyze and find out various chronic diseases based on various contributing parameters. This methodology of predicting chronic disease provides real-time assistance on diagnosis and prediction of the major chronic disease that is Diabetes. In this research data analysis and prediction, modeling is done by using the large datasets consisting of larger sets of different types of attributes. Research is done to identify which model works better, from the list of many machine learning algorithms; few algorithms were explored in the current paper. We are proposing a system that mainly works on five machine learning algorithms, those are, Naïve Bayes, Decision Tree, KNN Classifier, Random Forest, and SVM. The study of the prediction of chronic disease also involves the comparison of these different models to seek efficient results. Achieving high accuracy is an important aspect, based on the model that provides higher accuracy and efficacy the model will be selected as the final solution. In this research work, multiple machine learning algorithms are analyzed to predict the Chronic Diseases, among these analysis models Naïve Bayes will give the best results. In future work, this research would be extended to predict more such chronic disease and combine all of those models to create a better health prediction system. This Research can be applied in the medical field to diagnose and predict chronic diseases, this helps to reduce the dependency of medically trained professionals, and this will reduce the cost, effort, and time and enables the early detection of chronic diseases.