Visualization for Analyzing Usage Status from Dialogue Systems

Abstract

Dialogue systems, which give users quick and easy access to required information interactively, have been widely used in various fields. To increase user satisfaction, it is important to analyze usage status in detail after introduction of dialogue systems. However, since usage status is analyzed from a large number of logs and other collected information, it can take much time and effort to find beneficial information. This study proposes a method that utilizes the dialogue system introduced for public relations in universities to support the analysis of usage status by visualization, aiming discovery by intuitive awareness. In the results of evaluation using real data of the dialogue system, a variety of information that matches different purposes and unexpected information were discovered by intuitive awareness and supporting functions.