基于马尔可夫链的情感计算建模方法

Affective Computing Model Based on Markov Chain

定义了情感的两种状态及其两个基本的变化转移过程,并应用马尔可夫链构造了一个情感概率空间,建立模拟情感变化的情感模型,给出了情感能量、情感强度和情感熵等概念,用以描述情感特征与情感状态。通过Matlab的仿真计算,验证此模型可较好地模拟情感状态自发转移的动态过程,可用于情感机器人的情感模拟计算。为情感计算和情感自动生成理论研究提供了一种新途径。

The paper presents two kinds of states and two basic transferring processes in the affectivity, and builds an affective probability space applying Markov chain. Meanwhile an affective model is studied, which can simulate affective change. In this affective model, some concepts using for describing affective character and states are defined such as zero affective energy, affective intensity, affective entropy and so on. Simulations are done using Matlab software, and simulation results show that this kind of affective model can better simulate dynamic process of emotion transferring and change spontaneously. This model can be applied to affective robot.