METHOD FOR STATISTICALLY UNDERSTANDING VOICE LANGUAGE

摘要

The present invention provides a method for statistically understanding a voice language by using a statistical scheme without a manually tagged corpus, comprising a dialog act clustering step of clustering utterance sentences based on similar dialog acts; a named entities clustering step of extracting named entity candidate groups from the result of the dialog act clustering step and clustering the named entities based on surrounding context information of the extracted named entity candidate groups; and a main act clustering step of clustering main acts for each area based on the clustered dialog acts and named entities.;COPYRIGHT KIPO 2014