Advanced search
1 file | 1.36 MB Add to list

EvalNE : a framework for evaluating network embeddings on link prediction

Author
Organization
Abstract
In this paper, we present EvalNE, a Python toolbox for evaluating network embedding methods on link prediction tasks. Link prediction is one of the most popular choices for evaluating the quality of network embeddings. However, the complexity of this task requires a carefully designed evaluation pipeline to provide consistent, reproducible and comparable results. EvalNE simplifies this process by providing automation and abstraction of tasks such as hyper-parameter tuning and model validation, edge sampling and negative edge sampling, computation of edge embeddings from node embeddings, and evaluation metrics. The toolbox allows for the evaluation of any off-the-shelf embedding method without the need to write extra code. Moreover, it can also be used for evaluating link prediction methods and integrates several link prediction heuristics as baselines. Finally, demonstrating the usefulness of EvalNE in practice, we conduct an extensive analysis where we replicate the experimental sections of several influential papers in the community.
Keywords
graph embedding, link prediction, evaluation, reproducibility

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 1.36 MB

Citation

Please use this url to cite or link to this publication:

MLA
Mara, Alexandru Cristian. “EvalNE : a Framework for Evaluating Network Embeddings on Link Prediction.” Proceedings of 19th FEA Research Symposium. Ed. Kevin Dekemele. Ghent: Ghent University, 2019. 84. Print.
APA
Mara, A. C. (2019). EvalNE : a framework for evaluating network embeddings on link prediction. In K. Dekemele (Ed.), Proceedings of 19th FEA Research Symposium (p. 84). Presented at the 19th edition of the FEA Research Symposium, Ghent: Ghent University.
Chicago author-date
Mara, Alexandru Cristian. 2019. “EvalNE : a Framework for Evaluating Network Embeddings on Link Prediction.” In Proceedings of 19th FEA Research Symposium, ed. Kevin Dekemele, 84. Ghent: Ghent University.
Chicago author-date (all authors)
Mara, Alexandru Cristian. 2019. “EvalNE : a Framework for Evaluating Network Embeddings on Link Prediction.” In Proceedings of 19th FEA Research Symposium, ed. Kevin Dekemele, 84. Ghent: Ghent University.
Vancouver
1.
Mara AC. EvalNE : a framework for evaluating network embeddings on link prediction. In: Dekemele K, editor. Proceedings of 19th FEA Research Symposium. Ghent: Ghent University; 2019. p. 84.
IEEE
[1]
A. C. Mara, “EvalNE : a framework for evaluating network embeddings on link prediction,” in Proceedings of 19th FEA Research Symposium, iGent Tower, Tech Lane Ghent, 2019, p. 84.
@inproceedings{8611637,
  abstract     = {In this paper, we present EvalNE, a Python toolbox for evaluating network embedding methods on
link prediction tasks. Link prediction is one of the most popular choices for evaluating the quality of
network embeddings. However, the complexity of this task requires a carefully designed evaluation
pipeline to provide consistent, reproducible and comparable results. EvalNE simplifies this process
by providing automation and abstraction of tasks such as hyper-parameter tuning and model validation,
edge sampling and negative edge sampling, computation of edge embeddings from node
embeddings, and evaluation metrics. The toolbox allows for the evaluation of any off-the-shelf embedding
method without the need to write extra code. Moreover, it can also be used for evaluating
link prediction methods and integrates several link prediction heuristics as baselines. Finally, demonstrating
the usefulness of EvalNE in practice, we conduct an extensive analysis where we replicate
the experimental sections of several influential papers in the community.},
  author       = {Mara, Alexandru Cristian},
  booktitle    = {Proceedings of 19th FEA Research Symposium},
  editor       = {Dekemele, Kevin},
  keywords     = {graph embedding,link prediction,evaluation,reproducibility},
  language     = {eng},
  location     = {iGent Tower, Tech Lane Ghent},
  pages        = {1},
  publisher    = {Ghent University},
  title        = {EvalNE : a framework for evaluating network embeddings on link prediction},
  year         = {2019},
}