Development of Interactive Chatting Game Base on Sentiment Analysis Method

Abstract

In online game, storytelling is one of the most important factors that motivates game play. Experience of this narrative is delivered through various storytelling techniques [1]. Recently, there has been a lot of interest in developing embodied conversational agents (ECA) with multimodal interfaces [2]. Among multimodal interfaces, natural language processing is most promising techniques in commercial game development side. Traditionally in game industry, communication between player and virtual agent in game was conducted by selecting player's replay among 2~3 candidate sentences. But recent advances in natural language process enable player's sentence input in interactive drama game [3]. In this paper, we propose interactive chatting game based on sentiment analysis of player's input sentence. In order to understand player's emotions, we construct web-query based sentiment databases, and uses it with similarity ccomparison method. Our result could show that emergent narrative enables re-playable experience with natural language process technique in commercial online game.