Diagnosis of chronic disease in a predictive model using machine learning algorithm

Today, digitization in healthcare industry takes the advantage on advancements in clinical healthcare services. The extensive growth in data for monitoring and analyzing the patients outcomes in predicting and diagnosis of chronic diseases lacks in traditional methods and are replaced by technologies to gather the most relevant insights from the medical data by using predictive analytics with very useful tool of machine learning. The importance of using machine learning algorithms in the model for diagnosis, shows its ability in high classification accuracy rate in reduced computational time. In this paper, a study of various machine learning techniques are used in classification of chronic diseases like heart, kidney, diabetes and cancer from multiple dataset by reducing the dimensionality using feature selection. Feature selection plays a significant role in machine learning by selecting the critical features for diagnosing chronic diseases. The performance of the classifiers are evaluated based on several metrics like classification accuracy, sensitivity, specificity, precision, F1- measure, AUC (the area under the receiver operating characteristic (ROC) curve) criterion, and processing time.