Data and text mining from online reviews

Abstract

This paper reports on a thorough analysis of the scientific literature using data and text mining to uncover knowledge from online reviews due to their importance as user-generated content. In this context, more than 12,000 papers were extracted from publications indexed in the Scopus database within the last 15 years. Regarding the type of data, most previous studies focused on qualitative textual data to perform their analysis, with fewer looking for quantitative scores and/or characterizing reviewer profiles. In terms of application domains, information management and technology, e-commerce, and tourism stand out. It is also clear that other areas of potentially valuable applications should be addressed in future research, such as arts and education, as well as more interdisciplinary approaches, namely in the spectrum of the social sciences. This article is categorized under: Algorithmic Development > Text Mining Application Areas > Business and Industry.