Music Content Ranking using Affective Information on Social Media
R. Arun Kumar · International Journal for Research in Applied Science and Engineering Technology · 2019
It is easy to determine a person's emotion state from their social network presence online. But it is only easy for us humans to correlate these emotions. But in the field of music information retrieval and recommendation, emotion is considered contextual information that is hard to capture, albeit highly influential.In this study, we exploit the user's online presence in twitter to determine his/her state of mind and ultimately provide music suggestions.Particularly, we perform a large-scale study based on data sets containing now playing tweets.We extract affective contextual information from hashtags that are present in social media posts by applying an unsupervised sentiment dictionary approach.Subsequently, we utilize a state-of-the-art network embedding method to learn latent feature representations of users, tracks and hashtags.We find that the suitable ranking method helps in discerning the music preference of the user.For capturing context-specific preferences which is a more complex and personal task, we find that affective information and leveraging hashtags as context information are the best ranking strategies that outperform the other ranking strategies.