An information-pattern-based approach to novelty detection

ct 9 In this paper, a new novelty detection approach based on the identification of sentence level information patterns is 10 proposed. First, ``novelty'' is redefined based on the proposed information patterns, and several different types of infor11 mation patterns are given corresponding to different types of users' information needs. Second, a thorough analysis of sen12 tence level information patterns is elaborated using data from the TREC novelty tracks, including sentence lengths, named 13 entities (NEs), and sentence level opinion patterns. Finally, a unified information-pattern-based approach to novelty detec14 tion (ip-BAND) is presented for both specific NE topics and more general topics. Experiments on novelty detection on 15 data from the TREC 2002, 2003 and 2004 novelty tracks show that the proposed approach significantly improves the 16 performance of novelty detection in terms of precision at top ranks. Future research directions are suggested. 17

• Type: Journal
• Year: 2008
• Where: IPM
• Authors: Xiaoyan Li, W. Bruce Croft
• Keywords: Information Patterns, IPM 2008, Level Information Patterns, Sentence Level