Some experiments with the Dutch collection

Abstract

We performed some basic monolingual Dutch and bilingual English-Dutch experiments. The retrieval approach is very basic, without stemming or decompounding, using only a simple language model to rank the documents. In the bilingual task, the English queries have been analyzed by a system for question answering. The resulting queries are translated by dictionary lookup and ranked by the same basic retrieval system used in the monolingual task. © Springer-Verlag Berlin Heidelberg 2003.