When to cross Over? Cross-language linking using Wikipedia for VideoCLEF 2009

摘要

We describe Dublin City University (DCU)'s participation in the VideoCLEF 2009 Linking Task. Two approaches were implemented using the Lemur information retrieval toolkit. Both approaches rst extracted a search query from the transcriptions of the Dutch TV broadcasts. One method rst performed search on a Dutch Wikipedia archive, then followed links to corresponding pages in the English Wikipedia. The other method rst translated the extracted query using machine translation and then searched the English Wikipedia collection directly. We found that using the original Dutch transcription query for searching the\udDutch Wikipedia yielded better results.