Back

Predicting Latent Links from Incomplete Network Data Using Exponential Random Graph Model with Outcome Misclassification

Wu, Q.; Chen, S.

2019-11-25 neuroscience
10.1101/852798 bioRxiv
Show abstract

SO_SCPLOWUMMARYC_SCPLOWLink prediction is a fundamental problem in network analysis. In a complex network, links can be unreported and/or under detection limits due to heterogeneous noises and technical challenges during data collection. The incomplete network data can lead to an inaccurate inference of network based data analysis. We propose a new link prediction model that builds on the exponential random graph model (ERGM) by considering latent links as misclassified binary outcomes. We develop new algorithms to optimize model parameters and yield robust predictions of unobserved links. The new method is applied to a partially observed social network data and incomplete brain network data. The results demonstrate that our method outperforms the existing latent-contact prediction methods.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.