Topic modeling and its applications in materials science and engineering

Abstract

Topic modeling is an unsupervised task which helps to capture hidden semantics structure of words in a document. It is significant area of research due to its diverse applications in multidisciplinary fields such as materials sciences, chemical engineering, industrial aspects, social media analysis, bioinformatics and many more. In materials science and engineering, with the help of topic modeling methods, clustering of keywords related to material synthesis steps can be performed which further leads to the detection of topics related to various steps. e.g. keywords related to synthesis procedures can be clustered into topics such as centrifuging, grinding, heating, and dissolving. Classification technique may be used on these topics to make groups of materials based upon their synthesis steps such as hydrothermal, or solid-state. In this paper, an overview and evolution of prominent topic modeling techniques is presented. Also, applications of topic modeling in materials science are discussed.