Identification of Interrogatives Depending on User Responses and its Effects

Abstract

In recent years, with the development of natural language processing technology, there has been a growing interest in dialogue systems. However, at present, it cannot be said that the system and people are sufficiently able to have a natural dialogue. In this study, we propose an approach to question generation that aims to develop natural dialogues. By extracting the related words from utterances and selecting an interrogative word, a question is generated considering the context. When ``What'' or ``How'' is used in the question, the user's answer determines which is the interrogative word. The experimental results show the effectiveness of our proposed method of generating questions using the related words and determining interrogative words.