Application of machine learning techniques in chronic disease literature: from citation mapping to research front

This study aims to conduct a hybrid review on applying machine learning techniques in chronic disease literature using both bibliometric and systematic review techniques. The dataset consists of 206 Scopus indexed journal articles from 2004 to 2020. The bibliometric results identify the most contributing authors, journal sources, author network, bibliometric coupling of documents, and the co-citation network. The systematic review reveals the most promising research areas, which include machine learning algorithms integrated with other techniques such as deep learning, artificial neural network, and data mining to predict chronic diseases in gastroenterology, cardiology, and neurology. Although machine learning techniques are rising in popularity in chronic disease literature, there is more room for improvement such as the challenges involved in using machine learning to predict chronic diseases, feasibility studies, and the necessity of rehabilitation and readmission in hospitals to predict a chronic attack.