Citation-Based Document Categorization: An Approach Using Artificial Neural Networks

Magali Rezende Gouvêa Meireles, Beatriz Valadares Cendón, Qualitative, Quantitative Methods in Libraries, Feb 12, 2015
Commentary by Stephen Downes
files/images/Categorization_Maps.PNG

Meaning, said Ludwig Wittgenstein, is use. So the meaning of an academic paper would be how it is used, and the most common use of a paper is represented via citations. So instead of categorizing papers through their contents, which mixes together the useful and the useless, a categorization system based on citations might be more interesting. This paper uses neural networks to identify use-based clusters by drawing on citation data, which results in new categorizations distinct from what we might find in content based categorizations using titles and keywords. Good stuff, clearly written, part of the just-released special issue of Qualitative and Quantitative Methods in Libraries focusing on bibliometrics.

Views: 0 today, 110 total (since January 1, 2017).[Direct Link]