Doctoral thesis

Social shuffle : music discovery with tag navigation and social diffusion


92 p

Thèse de doctorat: Università della Svizzera italiana, 2011

English This thesis tackles the problem of discovering music for users in a social network, introducing the concept of social shuffle and its implementation as a live experiment in social based recommendation, Starnet, and show that recommendations based on a user’s social network is strongly effective in introducing a user to new music that she enjoys. I investigate the generation of tag clouds using Bayesian models and test the hypothesis that social network information is better than overall popularity for ranking new and relevant information. I propose three tag cloud generation models based on popularity, topics and social structure. I conducted two user evaluations to compare the models for search and recommendation of music with social network data gathered from Our survey shows that search with tag clouds is not practical whereas recommendation is promising. I report statistical results and compare the performance of the models in generating tag clouds that lead users to discover songs that they liked and were new to them. I find statistically significant evidence at 5% confidence level that the topic and social models outperform the popular model. I report on an experiment on social diffusion for music discovery. I describe the experimental methodology which includes the making of a music videos dataset and the creation of a social application. I give a statistical analysis of the participants ratings which shows that social diffusion leads to more good recommendations. I conclude and show that the social shuffle is an effective mechanism for information recommendation and that social relationships are relevant data to enhance information navigation.
  • English
Computer science and technology
License undefined
Persistent URL

Document views: 111 File downloads:
  • 2011INFO009.pdf: 74