A proposal of topic map based Chatterbot for non-English natural language input

In this paper, we propose a Chatterbot that returns an answer relevant to a question submitted by a user. We focus on non-English language, whose structure is more flexible. This requires a new framework to analyze the question. Our method divides the question sentence into triples, each of which is a combination of a noun, a particle, and a predicate. Expecting intelligent response to the question, we introduced the topic map which associates the units to suitable answers. In order to absorb the variant of particle usage, we applied cluster analysis to questions in Yahoo! Answers to group particles.