Evaluating Link Prediction Accuracy on Dynamic Networks with Added and Removed Edges
Ruthwik R. Junuthula,Kevin S. Xu,Vijay K. Devabhaktuni
DOI: https://doi.org/10.48550/arXiv.1607.07330
2016-07-26
Abstract:The task of predicting future relationships in a social network, known as link prediction, has been studied extensively in the literature. Many link prediction methods have been proposed, ranging from common neighbors to probabilistic models. Recent work by Yang et al. has highlighted several challenges in evaluating link prediction accuracy. In dynamic networks where edges are both added and removed over time, the link prediction problem is more complex and involves predicting both newly added and newly removed edges. This results in new challenges in the evaluation of dynamic link prediction methods, and the recommendations provided by Yang et al. are no longer applicable, because they do not address edge removal. In this paper, we investigate several metrics currently used for evaluating accuracies of dynamic link prediction methods and demonstrate why they can be misleading in many cases. We provide several recommendations on evaluating dynamic link prediction accuracy, including separation into two categories of evaluation. Finally we propose a unified metric to characterize link prediction accuracy effectively using a single number.
Social and Information Networks,Machine Learning,Physics and Society,Methodology