A probabilistic approach to personalized tag recommendation

• Main Authors: HU, Meiqun, LIM, Ee Peng, JIANG, Jing
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2010
• Subjects: Social tagging; Recommendation methods; Neighbor-based translation method; Databases and Information Systems; Numerical Analysis and Scientific Computing
• Online Access: https://ink.library.smu.edu.sg/sis_research/619; https://ink.library.smu.edu.sg/context/sis_research/article/1618/viewcontent/auto_convert.pdf
• Institution: Singapore Management University

In this work, we study the task of personalized tag recommendation in social tagging systems. To reach out to tags beyond the existing vocabularies of the query resource and of the query user, we examine recommendation methods that are based on personomy translation, and propose a probabilistic framework for incorporating translations by similar users (neighbors). We propose to use distributional divergence to measure the similarity between users in the context of personomy translation, and examine two variations of such similarity measures. We evaluate the proposed framework on a benchmark dataset collected from BibSonomy, and compare with personomy translation methods based on the query user solely and collaborative filtering. Our experimental results show that our neighbor based translation methods outperform these baseline methods significantly. Moreover, we show that the translations borrowed from neighbors indeed help ranking relevant tags higher than that based solely on the query user.