[3Xin2-40] Engagement-motivated Goal-Oriented Dialogue Dataset for Sales Talk
Keywords:Dialogue corpus, Engagement, Sales talk, Crowdsourcing
In this research, we propose a new concept of Engagement-motivated Goal-Oriented Dialogue (EGOD), a dialogue task designed to conversational objectives while incorporating three facets of user engagement: (1) willingness to continue a dialogue, (2) willingness to provide information, and (3) willingness to accept a dialogue goal.
Furthermore, we built a sales EGOD dataset for product sales and analyzed the dataset in order to realize an EGOD system.
We have constructed a high quality sales talk data set, as we confirm increased to user’s purchase intent in 58/98 dialogues.
Our analysis of the dataset suggests that avoiding utterances that decrease engagement is more important than making utterances that increase engagement in order to achieve the dialogue goal of increasing purchase intent.
Furthermore, we built a sales EGOD dataset for product sales and analyzed the dataset in order to realize an EGOD system.
We have constructed a high quality sales talk data set, as we confirm increased to user’s purchase intent in 58/98 dialogues.
Our analysis of the dataset suggests that avoiding utterances that decrease engagement is more important than making utterances that increase engagement in order to achieve the dialogue goal of increasing purchase intent.
Authentication for paper PDF access
A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.