LAUSR

BACKGROUND Semantic analyses of patents have been used for years to unlock technical knowledge. Nevertheless, information retrievable from patents remains widely unconsidered when making strategic decisions, when recruiting candidates or deciding which qualifications to offer to employees in technological fields. OBJECTIVES This paper provides an approach to evaluate whether competencies and competence demands in technological fields can be derived from patents and if this process can be automated to a certain extent. METHODS A sample of significant patents is analyzed with regard to comprised competence data via semantic structures like n-gram and Subject--Action-Object (SAO) analysis. The retrieved data is cleansed and matched semantically to inventor competencies from social career networks and checked for similarities. RESULTS A social career network profile analysis of significant inventors revealed a total of 570 competencies that were matched with the results of the n-gram and SAO analysis. Overall, 15%of the extracted social career network competence data were covered through extracted n-grams (87 out of 570 terms), while the SAO analysis showed a match rate of 18.8%, covering 107 terms. CONCLUSIONS The outlined approach suggests a partly automatable process of promising character to identify technological competence demands in patents.