Micro science and technology fields requiring mathematically trained contributors: Topic modeling using journal paper abstracts

A shortage of mathematically trained individuals who can integrate mathematical knowledge with other types of knowledge and skills and thereby contribute to innovation has recently become apparent. Finding clues to help in solving this problem requires a) identification of scientific and technology fields (STFs) that require mathematically trained contributors and b) identification of higher education institutions (HEIs) that provide students an opportunity to develop the mathematical knowledge needed in such STFs. Previous work on this problem has shown that the granularity of STFs is coarse, such as at the information theory level. An in-depth discussion of the educational practices, e.g., project-based learning and industrial internship programs, at mathematical HEIs requires analysis at the micro-STF level. To enable such analysis, a method is presented for discovering topics hidden in a collection of journal paper abstracts. For example, topic modeling using the latent Dirichlet allocation algorithm implemented in Python enabled the discovery of the topics covered studied at the three highest-ranked mathematical HEIs in the information theory field.