Research On A Layered Model Of Affect

Emotion is an important part of human intelligence. At present, with the rapid development of computer technology, artificial methods and techniques which are used to imitate and extend human emotions, have broad prospects and potential economic values in human-computer interaction, anthropomorphic robotics, education and entertainment, which can make machines have the ability of recognizing, understanding and expressing emotions. They have become a hot topic of academic circles and industrial circles.Emotions which are influenced by many factors are changeable, complex and ambiguous, so making computers intelligent is a topic full of challenge. How to establish a proper mathematical model of affect is a key problem that should be solved. At present, many models have been proposed, but they just took several emotions, which changed in the small scope of emotional range, into account to simplify models, and they ignored external affective stimulus, personality influence, emotion decay, emotion interactions, and mood influence and so on. So these models could not describe the complex processes of emotions change, and lose practicality.Therefore, to improve the existing affective models, this paper builds a new layered model of affect, taking external factors and interior factors into account:1)Emotion decayEmotions decay differently with different personality, and for one person, different emotions decay differently. Different decay functions are built for different emotion, according to personality.2)Mood influenceA person has different reaction to the same external stimulus with different mood. a renewal equation of mood is built, taking personality influence, external stimulus and mood decay into account. the mood influence is obtained using transformation relations between emoiton and mood3)Emotion interactionsEmotion interactions are obtained on the basis of our group`s research.4)External stimulusIn the experiment, external stimulus are a large number of materials marked with emotions.To show the efficiency of the model, an emotional prediction system is designed and implemented on the basis of the model. After 80 students` tests, the system gets good results with 80 percent accuracy, which shows that the model is efficient and practical. Doing some basic research for harmonious human-computer interactions and emotional robo, the model has some academic significance and application values.