Although interesting results and findings occurred from this study, limitations do exist. (1) The keyword structure was
designed through a number of trials to ensure the most effective and feasible search space. However, there may still be some
related works that this keyword structure has not captured. (2) The categorizations and titles of both the actual modeling
categories and generative studies are not necessarily homogenous. Multiple and different topical areas exist in some areas
and a more nuanced, sub-categorization evaluation may provide ample room for identification of sub-clusters or variations
in interpretations. (3) The Scopus database is relatively comprehensive and has certain advantages as a database, but is not
as encompassing in capturing information as Google Scholar. Not all journals that are on the listing appeared in Scopus for all
years. Thus, some early publications may have been missed. Also, Scopus may have limited electronic information before the
1978 time period, potentially limiting pre-1978 articles that may have been captured in older electronic systems. (4) The
literature mapping and network analysis methodology presented in this paper shows how a subject area can be objectively
reviewed to identify the major articles and investigators. However, the methodology is not able to interpret the knowledge in
these articles to explore the reasons why particular articles have been central to the development of the field. Future review
efforts can focus on the development of tools and methodologies to address these limitations