A Context-Based Talker
一种基于语境的对话模型

In order to make computer understand human languages much better and achieve certain tasks according to human intentions, it is required to carefully model the elementary concepts and the underlying theoretical principles for driving a conversation generating and developing. Under the given linguistic circumstances, to some extent, it might be hopeful to build up a computational model which can predict and control the sustaining development of conversational activity. This paper is going to make a formalized primary discussion on contextual features which influence conversational activity, design an algorithm processing ellipses, and build a Context-Based Talker (CBT). The empirical analysis shows, to some extent, CBT can control and guarantee the coherence and natural development of conversational activity.