Segmented Spoken Document Retrieval Using Word Co-occurrence Information.
Hara Kensuke, Hiroaki Taguchi, K. Nakajima, Masanori Takehara, Satoshi Tamura, Satoru Hayamizu · 2014
This paper shows several approaches for NTCIR-11 Spo-kenQuery&Doc [1]. This paper proposes several schemes to use word co-occurrence information for spoken document re-trieval. Automatic transcriptions of spoken documents usu-ally contain mis-recognized words, making the performance of spoken document retrieval signicantly decrease. The co-sine similarity to measure a document similarity must be investigated for spoken documents. It is also dicult to re-trieve a segmented document having few terms. To cope with these problem, we utilize Pointwise Mutual Informa-tion (PMI). We compute a recognition condence for each term appeared in a transcription to drop mis-recognized words. We also investigate a PMI-based document com-parison approach. Furthermore, a segmented-document re-trieval method is also proposed. Experiments were con-ducted to evaluate these methods using NTCIR-11 test sets.