Emotion Analysis Using Multimodal Data

As part of the various treatments and explorations of massive data on the web, emotion modelling is part of a method that aims to understand and predict users' emotional states. Behavioral analysis from multimedia content enables the detection of reactions and attitudes elicited by the music and lyrics of a song. This paper presents the development of a Deep Learning-based model for analyzing listeners' emotions from multimodal data, which also aims to optimize decision-making based on these reactions. Efficient classification of emotions will contribute to the optimization of recommendation systems.