Author Clustering Using SPATIUM

摘要

This paper presents the author clustering problem and compares it to related authorship attribution questions. The proposed model is based on a distance measure called Spatium derived from the Canberra measure (weighted version of L norm). The selected features consist of the 200 most frequent words and punctuation symbols. An evaluation methodology is presented and the test collections are extracted from the PAN CLEF 2016 evaluation campaign. In addition to those, we also consider two additional corpora reflecting the literature domain more closely. Based on four different languages, the evaluation measures demonstrate a high precision and F1 for all 20 test collections. A more detailed analysis provides reasons explaining some of the failures of the Spatium model.