Journal article mining: The scholarly publishers' perspective

Abstract

The essence of text mining and data mining is that a machine and software are used for content analysis of large digital corpora. The Publishing Research Consortium commissioned a study on content mining of scholarly journal articles with 29 expert interviews and an international survey among publishers. The main results are: (i) content mining developments appear to be accelerating with more applications in more areas; (ii) third-party demand for content mining is widespread but still at low levels of frequency; (iii) publishers'permissions for content mining are quite liberal, especially for research-driven mining requests; (iv) half of the publisher respondents undertake mining of their own content; and (v) content mining is on the rise - publishers and third parties both report an increase in planned mining activities. As content mining of journal articles spreads and intensifies, cross-publisher solutions can better help facilitate content mining. The study investigated the interest and willingness of publishers to support a set of different solutions, from one shared content mining platform to commonly agreed access terms for mining and standardization of mining-friendly content formats. © Eefke Smit and Maurits van der Graaf 2012.