Skip to main content

Table 4 Comparison of DS detection performances for the proposed approach using the verification data, the proposed approach with/without sentence clustering, and manual transcription

From: Miscommunication handling in spoken dialog systems based on error-aware dialog state detection

Average Detection Rate (%)
  Keyword spotting PA − SC + VD PA − SC PA + SC Manual transcription
Base DS 50.16 65.30 75.39 83.95 94.57
Mis-U DS 51.96 72.59 80.68 81.46 92.69
  1. PASC + VD proposed approach without sentence clustering using the verification data, PASC proposed approach without sentence clustering, PA + SC proposed approach with sentence clustering