Homogeneous Markov Chain for Modeling Emotional Interactions

In this paper, we develop finite state models of emotional interactions based on homogeneous Markov chain: an emotional agent takes a state from a finite set of basic emotion-happiness, surprise, fear, anger, disgust and sadness. And the emotional agent updates its state depending on its current state and a state of its input (also a state-emotion). We classify six basic emotions to three big classes: positive, neutral, and negative emotion. Then, we associate any two big classes to gain the results of interactions. The results of interactions are obtained by calculating the value of emotional valence and arousal. Finally, emotions poly-tropism is confirmed and the temporal length of emotional state is calculated. All these results are basing on matrix of transition probability and the features of homogeneous Markov chain by experiments.