WiKNN – a tool for thematic text classification


WiKNN (Wikipedia K-Nearest Neighbors) is an online text classifier service for Polish and English. It provides hierarchically labelled classification of user-submitted texts with Wikipedia categories. The tool automatically assigns thematic classifiers from Wikipedia to Polish and English texts. This allows for indexing of text collections with categories from Polish and English Wikipedia. WiKNN is available through a web-based interface at: http://pelcra.clarin-pl.eu/tools/classifier/and and as a REST service with interactive documentation at: http://clarin.pelcra.pl/apidocs/wiknn.

Bibliographic address of the main publication (in case of using Chronocorpus, please cite this publication):

Pęzik, P., & Buczek, M. (2015). Druga wersja klasyfikatora tematycznego tekstów WiKNN. Zadanie A23. Punkt kontrolny M16. http://pelcra.clarin-pl.eu/tools/pdf/wiknn.pdf

Examples of applications

Klasyfikator został użyty do anotacji kategoriami Wikipedii korpusu NKJP: http://pelcra.clarin-pl.eu/NKJP/.