An artificial intelligence algorithm analyzing 30 years of research in mass appraisals

The research papers issued in scientific journals, for a variety of thematic areas, are not only increasing, nonetheless exhibit an exponential growth over the last years. Accordingly, the researchers, struggle to retrieve information apropos of novel knowledge and get informed in their field, while the rigor and at the same time, the extensive composition of surveys, reviews, and overviews of research works, has become difficult or even impossible, as the number of the available research studies is enormous. However, such reviews, contain vital information regarding the evolution of a scientific subject, the trends of the literature, the most significant concepts, and the concealed associations among research papers, their references, as well as authors’ clusters. In this work, a scientometric study of the relevant to Mass Appraisals literature is for a first time accomplished, regarding the numerical models, computational procedures, and automated methods, utilized in the Mass Appraisals and Property Valuations literature. The study is based on an adequate pool of papers, constituted in Scopus database, utilizing a machine learning algorithm developed from one of the authors, for multidimensional scaling and clustering of the keywords found in the papers’ database, the authors and their cooperation and the co-occurrences of the references in the papers studied. The time-series of the most frequent keywords are also computed, demonstrating the evolution of the mass appraisals research and identifying future trends.