An Approach for Sentiment Classification of Music
Francesco Colace, Luca Casaburi · 2016
In recent years, the music recommendation systems and dynamic generation of playlists have become extremely promising research areas. Thanks to the widespread use of the Internet, users can store a consistent set of music data and use them in the everyday context thanks to portable music players. The problem of modern music recommendation systems is how to process this large amount of data and extract meaningful content descriptors. The aim of this paper is to compare different approaches to decode the content within the mood of a song and to propose a new set of features to be considered for classification.