Exploring the Importance of Individual Differences to the Automatic Estimation of Emotions Induced by Music

Abstract

The goal of this study was to evaluate the impact of the inclusion of listener-related factors (individual differences) on the prediction of music induced affect. A group of 24 subjects listened to a set of music excerpts previously demonstrated to express specific emotional characteristics (in terms of Arousal and Valence), and we collected information related to listeners' stable (personality, emotional intelligence, attentiveness, music preferences) and transient (mood, and physiological activity) states. Through a series of regression analysis we identified those factors which have a significant explanatory power over the affective states induced in the listeners. Our results show that incorporating information related to individual differences permits to identify more accurately the affective states induced in the listeners, which differ from those expressed by the music.