Using Non-topological Node Attributes to Improve Results of Link Prediction in Social Networks

Yu Zhang, Feng Li, Bin Xu, Kening Gao, Ge YuWISA 2012 · OpenReview

Abstract

This paper examines the importance of non-topological node attributes for link prediction in social networks. Rank method and supervised learning method were introduced to show the role of the node attributes in link prediction respectively. A rule for choosing the appropriate node attributes was discussed and a method for aggregating two node attributes was proposed. The result of the experiments on a blog dataset showed that using non-topological node attributes make a better performance in link prediction.
BibTeX
@inproceedings{zhang2012using,
  author = {Yu Zhang and Feng Li and Bin Xu and Kening Gao and Ge Yu},
  title = {{Using Non-topological Node Attributes to Improve Results of Link Prediction in Social Networks}},
  booktitle = {WISA},
  year = {2012},
  url = {https://openreview.net/forum?id=9HrbX6yQFY},
}

Loading related content...

Figure