Estimating network edge probabilities by neighborhood smoothing
arXiv:1509.08588
Abstract
The estimation of probabilities of network edges from the observed adjacency matrix has important applications to predicting missing links and network denoising. It has usually been addressed by estimating the graphon, a function that determines the matrix of edge probabilities, but this is ill-defined without strong assumptions on the network structure. Here we propose a novel computationally efficient method, based on neighborhood smoothing to estimate the expectation of the adjacency matrix directly, without making the structural assumptions that graphon estimation requires. The neighborhood smoothing method requires little tuning, has a competitive mean-squared error rate, and outperforms many benchmark methods on link prediction in simulated and real networks.
22 pages, 4 figures, 3 table
References in corpus (4)
Cited by in corpus (13)
- On Robustness of Principal Component Regression
- Network cross-validation by edge sampling
- Iterative Collaborative Filtering for Sparse Matrix Estimation
- Generalized linear models with low rank effects for network data
- Link prediction for egocentrically sampled networks
- EM-Based Smooth Graphon Estimation Using Bayesian and Spline-Based Approaches
- Reducing Crowdsourcing to Graphon Estimation, Statistically
- Optimal link prediction with matrix logistic regression
- Revisiting Spectral Graph Clustering with Generative Community Models
- Graphon estimation via nearest neighbor algorithm and 2D fused lasso denoising
- Deconvolution with Unknown Error Distribution Interpreted as Blind Isotonic Regression
- Distributed Cartesian Power Graph Segmentation for Graphon Estimation
- On the Estimation of Network Complexity: Dimension of Graphons