User-to-User Recommendations in a Fashion Portal, Utilizing a Social Network, Implicit Feedback and Clustering Approaches
Authors
Term
4. term
Education
Publication year
2015
Submitted on
2015-06-03
Pages
108
Abstract
In this report we address the social link prediction problem, applying our research to design a recommender system for the Sobazaar social network. This domain shows several interesting characteristics. Firstly, differently from most recommender system scenarios, solely Implicit Feedback is available. Secondly, the evolution of the fashion trends causes complex social dynamics to appear in the structures of the network, therefore finding communities of users with similar characteristics is crucial to understand the users' behavior. We conclude that by combining clustering with information that can be extracted from the users' activity it is possible to positively contribute to the social link problem.
Keywords
Documents