Latent Semantic Analysis Applied to Tech Mining

Abstract

This paper presents an approach to bibliometric analysis in the context of technology mining. Bibliometric analysis refers to the use of publication database statistics, e.g., hit counts relevant to a topic of interest. Technology mining facilitates the identification of a technology's research landscape. Our contribution to bibliometrics in this context is the use of a technique known as Latent Semantic Analysis (LSA) to reveal the concepts that underlie the terms relevant to a field. Using this technique, we can analyze coherent concepts, rather than individual terms. This can lead to more useful results from our bibliometric analysis. We present results that demonstrate the ability of Latent Semantic Analysis to uncover the concepts underlying sets of key terms, used in a case study on the technologies of renewable energy.

Keywords: tech mining, latent semantic analysis