The Domain-Specific Track at CLEF 2007

The domain-specific track uses test collections from the social science domain to test monolingual and cross-language retrieval in structured bibliographic databases. Special attention is given to the existence of controlled vocabularies for content description and their potential usefulness in retrieval. Test collections and topics are provided in German, English and Russian. This year, a new English test collection e CSA Sociological Abstracts database) was added. We present an overview of the CLEF domain-specific track including a description of the tasks, collections, topic preparation, and relevance assessments as well as contributions to the track. A summary of results is given. The track participants experimented with different retrieval models ranging from classic vector-space to probabilistic to language models. The controlled vocabularies were used for query expansion or as bilingual dictionaries for query translation. Categories and Subject Descriptors H.3 [Information Stor...

• Type: Conference
• Year: 2007
• Where: CLEF
• Authors: Vivien Petras, Stefan Baerisch, Maximilian Stempfhuber
• Keywords: CLEF 2007, Controlled Vocabularies, Domain-specific Track, Information Retrieval, Test Collection