P06-1061 redundancy . Shown from Table 3 , the segment retrieval results have achieved high recall
P06-1061 extraction performance . Our current segment retrieval method requires the training
P06-1061 retrieval using HMMs We propose a segment retrieval approach for performing the first
P06-1061 have presented and evaluated our segment retrieval method . Document-wise retrieval
P06-1061 in Definition 4 to measure the segment retrieval accuracy . Definition 4 . Document-wise
P06-1061 process consists of two steps : a segment retrieval step followed by an extraction
P06-1061 evaluate average document-wise segment retrieval redundancy , as defined in Definition
P06-1061 learn an extraction-relevance segment retrieval system from the labelled texts
P06-1061 IE model , using the proposed segment retrieval approach , and comparing their
P06-1061 segment HMM IE system using the segment retrieval method has not only achieved
W05-0832 information that is useful in segment retrieval in the translation memory . For
P06-1061 performance Since the purpose of our segment retrieval is to identify relevant segments
P06-1061 observation sequences are used . 5.2 Segment retrieval using HMMs After a retrieval
P06-1061 on the goodness of a particular segment retrieval method for our purpose : the
P13-1018 precision or for recall . 3 Parallel Segment Retrieval We will first abstract from the
P06-1061 imposing an extraction-relevant text segment retrieval in the segment HMM IE modelling
P06-1061 segments retrieved ) by a given segment retrieval system . We consider two criteria
P06-1061 extractor HMM . 5.1 Training HMMs for segment retrieval To train a retrieval HMM , it
P06-1061 most correctly filtered by the segment retrieval system only when all the extraction
P06-1061 sequence models used for IE . Segment retrieval for extraction is an important
hide detail