Text Mining Using Latent Semantic Analysis: An Illustration through Examination of 30 Years of Research at JIS

Abstract

ABSTRACT Big Data presents a tremendous challenge for the accounting profession today. This challenge is characterized by, among other things, the explosive growth of unstructured data, such as text. In recent years, new text-mining methods have emerged to turn unstructured textual data into actionable information. A critical role of accounting information systems (AIS) research is to help the accounting profession assess and utilize these methodologies in an accounting context. This paper introduces the latent semantic analysis (LSA), a text-mining approach that discovers latent structures in unstructured textual data, to the AIS research community. An LSA-based approach is used to analyze AIS research as published in the Journal of Information Systems (JIS) over the last 30 years. JIS research serves as an appropriate domain of analysis because of a perceived need to contextualize the scope of AIS research. The research themes and trends resulting from this analysis contribute to a better understanding of this identity.