<?xml version="1.0" standalone="yes"?>
<Paper uid="H94-1034">
  <Title>Tagging Speech Repairs</Title>
  <Section position="3" start_page="0" end_page="187" type="intro">
    <SectionTitle>
2. Previous Work
</SectionTitle>
    <Paragraph position="0"> Several different strategies have been discussed in the literature for detecting and correcting speech repairs. One way to compare the effectiveness of these approaches is to look at their recall and precision rates. For detecting repairs, the recall rate is the number of correctly detected repairs compared to the number of repairs, and the precision rate is the number of detected repairs compared to the number of detections (including false positives). But the true measures of success are the correction rates. Correction recall is the number of repairs that were properly corrected compared to the number of repairs. Correction precision is the number of repairs that were properly corrected compared to the total number of corrections.</Paragraph>
    <Paragraph position="1"> One of the first computational approaches was that taken by  Hindle (I 983), who used a deterministic parser augmented with rules to look for matching categories and matching strings of words. Hindie achieved a correction recall rate of 97% on his corpus; however, this was olbtalned by assuming that speech repairs were marked by an explicit &amp;quot;edit signal&amp;quot; and with part-of-speech tags externally supplied. null The SRI g~up (Bear, Dowding and Shn%erg, 1992) removed the assumptiml of an explicit edit signal, and employed simple pattern matching techniques for detecting and correcting modification repairs (they removed all utterances with abridged repairs from their corpus). For detection, they were able to achieve a recall rate of 76%, and a precision of 62%, and they were able to find the correct repair 57% of the time, leading to an overall correction recall of 43% and correetion precision of 50%. They also tried combining syntactic and semantic knowledge in a &amp;quot;parser-first&amp;quot; approach--first try to parse the input and if that fails, invoke repair strategies based on their pattern matehing technique. In a test set of 756 utterances containing 26 repairs (Dowding et al., 1993), they obtained a detection recall rate of 42% and a precision of 84.6%; for correction, they obtained a recall rate of 30% and a precision rate of 62%.</Paragraph>
    <Paragraph position="2"> Nakatani and Hirschberg (1993) investigated using acoustic informarion to detect the interruption point of speech repairs. In their corpus, 74% of all repairs are marked by a word fragment. Using hand-transcribed prosodic annotations, they trained a classifier on a 172 utterance training set to identify the interruption point (each utterance contained at least one repair). On a test set of 186 utterantes containing 223 repairs, they obtained a recall rate of 83.4% and a precision of 93.9% in detecting speech repairs. The clues that they found relevant were duration of pause between words, presence of fragments, and lexical matching within a window of three words. However, they do not address the problem of determining the correction or distinguishing modification repairs from abridged repairs.</Paragraph>
  </Section>
class="xml-element"></Paper>