Reusable Research? A Case Study in Named Entity Recognition
Marieke van Erp, Lourens van der Meij · 2013
Named Entity Recognition (NER) is a crucial piece of knowl- edge acquisition infrastructure for the Semantic Web. Thus, being able to build-upon and reuse state-of-the art approaches in this domain is essential for continued progress. In this paper, we present results in at- tempting to reproduce the algorithms presented by Freire et al. in ESWC 2012 (1). Based on this result, we come to the conclusion that even if ex- periments are described in detail it is still very dicult to reproduce the experiments and pinpoint the particular diculties in this use case. Finally, we evaluate our attempts as well as those of (1) on the named entity recognition task.