• Corpus ID: 235683423

Edge Proposal Sets for Link Prediction

  title={Edge Proposal Sets for Link Prediction},
  author={Abhay Singh and Qian Huang and Sijia Huang and Omkar Bhalerao and Horace He and Ser-Nam Lim and Austin R. Benson},
Graphs are a common model for complex relational data such as social networks and protein interactions, and such data can evolve over time (e.g., new friendships) and be noisy (e.g., unmeasured interactions). Link prediction aims to predict future edges or infer missing edges in the graph, and has diverse applications in recommender systems, experimental design, and complex systems. Even though link prediction algorithms strongly depend on the set of edges in the graph, existing approaches… 

Figures and Tables from this paper

Citation network applications in a scientific co-authorship recommender system
This paper proposes a new pipeline that effectively utilizes citation data in the link prediction task on the co-authorship network and explores the capabilities of a recommender system based on data aggregation strategies on different graphs.
Improving random walk rankings with feature selection and imputation Science4cast competition, team Hash Brown
This paper details the model, its intuition, and the performance of its variations in the test set of the Science4cast Competition.


Link Prediction Based on Graph Neural Networks
A novel $\gamma$-decaying heuristic theory is developed that unifies a wide range of heuristics in a single framework, and proves that all these heuristic can be well approximated from local subgraphs.
A Survey of Link Prediction in Social Networks
This article surveys some representative link prediction methods by categorizing them by the type of models, largely considering three types of models: first, the traditional (non-Bayesian) models which extract a set of features to train a binary classification model, and second, the probabilistic approaches which model the joint-probability among the entities in a network by Bayesian graphical models.
Improving Network Community Structure with Link Prediction Ranking
Experimental results show that applying the approach to different networks can significantly refine community structure and observe that performance of link prediction ranking is correlated with certain network properties, such as the network size or average node degree.
Link Prediction in Complex Networks: A Survey
Link-Prediction Enhanced Consensus Clustering for Complex Networks
This work proposes a novel consensus clustering algorithm to enhance community detection on incomplete networks that utilizes existing community detection algorithms that process networks imputed by their link prediction based sampling algorithm and merges their multiple partitions into a final consensus output.
Representation Learning on Graphs: Methods and Applications
A conceptual review of key advancements in this area of representation learning on graphs, including matrix factorization-based methods, random-walk based algorithms, and graph neural networks are provided.
On the Bottleneck of Graph Neural Networks and its Practical Implications
It is shown that existing, extensively-tuned, GNN-based models suffer from over-squashing and that breaking the bottleneck improves state-of-the-art results without any hyperparameter tuning or additional weights.
Inductive Representation Learning on Large Graphs
GraphSAGE is presented, a general, inductive framework that leverages node feature information (e.g., text attributes) to efficiently generate node embeddings for previously unseen data and outperforms strong baselines on three inductive node-classification benchmarks.
Adaptive Universal Generalized PageRank Graph Neural Network
This work introduces a new Generalized PageRank (GPR) GNN architecture that adaptively learns the GPR weights so as to jointly optimize node feature and topological information extraction, regardless of the extent to which the node labels are homophilic or heterophilic.
The link-prediction problem for social networks
Experiments on large coauthorship networks suggest that information about future interactions can be extracted from network topology alone, and that fairly subtle measures for detecting node proximity can outperform more direct measures.