A ringtone recommendation agent based on a Bayesian model of user emotion
Yong-Jun Kim, Sung‐Bae Cho · 2009
This paper presents a ringtone recommendation agent system which utilizes user's emotion to recommend an appropriate ringtone and the suitable volume level of a mobile phone. The system uses a Bayesian network (BN) to infer user's emotion and the volume level at the current situation. After inferring user's emotion, a ringtone is selected from the pool of various kinds of genres of ringtones. Also, the volume level can be controlled according to the inferred volume level. User's feedback on the inferred values can be used to retrain the BN to provide the user with a personalized service. Experimental results showed that the proposed system is quite convenient and very useful.