Loading...
Towards Quantifying Sampling Bias in Network Inference
Title / Series / Name
Publication Volume
Publication Issue
Pages
Editors
Keywords
bias
relational classification
sampling networks
Computer Networks and Communications
Software
relational classification
sampling networks
Computer Networks and Communications
Software
URI
https://hdl.handle.net/20.500.14018/27367
Abstract
Relational inference leverages relationships between entities and links in a network to infer information about the network from a small sample. This method is often used when global information about the network is not available or difficult to obtain. However, how reliable is inference from a small labeled sample How should the network be sampled, and what effect does it have on inference error How does the structure of the network impact the sampling strategy We address these questions by systematically examining how network sampling strategy and sample size affect accuracy of relational inference in networks. To this end, we generate a family of synthetic networks where nodes have a binary attribute and a tunable level of homophily. As expected, we find that in heterophilic networks, we can obtain good accuracy when only small samples of the network are initially labeled, regardless of the sampling strategy. Surprisingly, this is not the case for homophilic networks, and sampling strategies that work well in heterophilic networks lead to large inference errors. This finding suggests that the impact of network structure on relational classification is more complex than previously thought.
Topic
Publisher
Place of Publication
Type
Conference paper
Date
2018-04-23
Language
ISBN
9781450356404
Identifiers
10.1145/3184558.3191567