Using linguistic features to find examples of conversational 'problem'
What is it about?
We explore ways in which linguistic features (apart from the meanings of words) can be used to identify places in which conversational exhibit "problems". We explore some of the ways that have been suggested and find that there are no simple unambiguous markers. However combinations of several markers may work to identify points at which conversations 'break down'.
Why is it important?
This will help those dealing with large data sets including those dealing with call centre data
The following have contributed to this page: David Zeitlyn
In partnership with: