Such collaborative lexicography projects as Wiktionary are becoming strong competitors for traditional semantic resources just as Wikipedia has already become for expert-built knowledge bases. Keeping the data obtained from the general public crowd in good quality is a very challenging problem because of the fuzzy nature of the crowdsourcing phenomena. The presented study focuses on predicting the word of the week articles on the Russian Wiktionary by treating this problem as a binary classification task. The best proposed model is based on the Naïve Bayes classifier and has weighted average precision, recall, and F1-measure values of 87% by evaluating on the provided dataset.
Dieser Datensatz wurde nicht während einer Tätigkeit an der Universität Mannheim veröffentlicht, dies ist eine Externe Publikation.
Search Authors in
