Enriching WordNet with folksonomies

Abstract

Manually constructed thesauri are not updated regularly, so they are hard to catch the fast emergence of new words. Moreover, the vocabularies of the professionals who construct the thesauri may not completely match the vocabularies of normal users. Recently, Folksonomy services are very popular and highly sensitive to information drift and the change of users' vocabularies. In this paper, we explore a method for enriching formal thesauri with informal Folksonomies. We demonstrate our method by semi-automatically enriching WordNet with new words emerging from a social bookmark service. Tags are related to each other by the subsumption relationships extracted from Folksonomies. New words are recommended to be placed in appropriate synsets of the WordNet hierarchy. An initial evaluation on our experimental result shows the effectiveness of our method. © 2008 Springer-Verlag Berlin Heidelberg.