Dropout Training of Matrix Factorization and Autoencoder for Link Prediction in Sparse Graphs
This work addresses link prediction in sparse graphs, which is important for network analysis applications, but it is incremental as it combines existing techniques with dropout regularization.
The paper tackled link prediction in sparse graphs by proposing MF+AE, a model that jointly trains matrix factorization and autoencoder with dropout to prevent overfitting, and showed it consistently outperforms competing methods on six real-world datasets, particularly for graphs with non-cohesive structures.
Matrix factorization (MF) and Autoencoder (AE) are among the most successful approaches of unsupervised learning. While MF based models have been extensively exploited in the graph modeling and link prediction literature, the AE family has not gained much attention. In this paper we investigate both MF and AE's application to the link prediction problem in sparse graphs. We show the connection between AE and MF from the perspective of multiview learning, and further propose MF+AE: a model training MF and AE jointly with shared parameters. We apply dropout to training both the MF and AE parts, and show that it can significantly prevent overfitting by acting as an adaptive regularization. We conduct experiments on six real world sparse graph datasets, and show that MF+AE consistently outperforms the competing methods, especially on datasets that demonstrate strong non-cohesive structures.