Link Prediction in Signed Networks
Signed networks represent the real world relationships, which are both positive or negative. Recent research works focus on either discriminative or generative based models for signed network embedding. In this paper, we propose a generative adversarial network (GAN) model for signed network which unifies generative and discriminative models to generate the node embedding. Our experimental evaluations on several datasets, like Slashdot, Epinions, Reddit, Bitcoin and Wiki-RFA indicates that the proposed approach ensures better macro F1-score than the existing state-of-the-art approaches in link prediction and handling of sparsity of signed networks.- doi
- 10.1145/3372923.3404805
- name
- Link Prediction in Signed Networks
- pages
- 2
- acm_url
- https://dl.acm.org/doi/10.1145/3372923.3404805
- authors
- Roshni Chakraborty, Ritwika Das, Nilotpal Chakraborty
- doi_url
- https://doi.org/10.1145/3372923.3404805
- license
- restricted
- summary
- Signed networks represent the real world relationships, which are both positive or negative. Recent research works focus on either discriminative or generative based models for signed network embedding. In this paper, we propose a generative adversarial network (GAN) model for signed network which unifies generative and discriminative models to generate the node embedding. Our experimental evaluations on several datasets, like Slashdot, Epinions, Reddit, Bitcoin and Wiki-RFA indicates that the proposed approach ensures better macro F1-score than the existing state-of-the-art approaches in link prediction and handling of sparsity of signed networks.
- keywords
- Signed network, link prediction, generative adversarial network, structural balance theory
- source_pdf
- HT-2020_51-35_3372923/3372923.3404805.pdf
- import_kind
- full_text
- open_access
- false
- ccs_concepts
- (empty)
- displayAuthor
- Roshni Chakraborty, Ritwika Das, Nilotpal Chakraborty
Powered by Seed HypermediaOpen App