Dublin City University at CLEF 2006: Cross-Language Speech Retrieval (CL-SR) Experiments

The Dublin City University participation in the CLEF 2006 CL-SR task concentrated on exploring the combination of the multiple fields associated with the documents. This was based on use of the extended BM25F field combination model originally developed for multi-field text documents. Additionally, we again conducted runs with our existing information retrieval methods based on the Okapi model. This latter method required an approach to determining approximate sentence boundaries within the free-flowing automatic transcription provided to enable us to use our summarybased pseudo relevance feedback (PRF). Experiments were conducted only for the English document collection. Topics were translated into English using Systran V3.0 machine translation.