Network Model Selection for Task-Focused Attributed Network Inference

TL;DR

This paper introduces a modular methodology for selecting the most suitable network models for specific predictive tasks, demonstrating significant performance improvements across multiple online user activity datasets.

Contribution

It proposes a general, interpretable framework for network model selection tailored to task-specific evaluation, enhancing predictive accuracy.

Findings

Model selection significantly improves task performance.

Methodology is effective across diverse datasets.

Order of magnitude performance increase observed.

Abstract

Networks are models representing relationships between entities. Often these relationships are explicitly given, or we must learn a representation which generalizes and predicts observed behavior in underlying individual data (e.g. attributes or labels). Whether given or inferred, choosing the best representation affects subsequent tasks and questions on the network. This work focuses on model selection to evaluate network representations from data, focusing on fundamental predictive tasks on networks. We present a modular methodology using general, interpretable network models, task neighborhood functions found across domains, and several criteria for robust model selection. We demonstrate our methodology on three online user activity datasets and show that network model selection for the appropriate network task vs. an alternate task increases performance by an order of magnitude in our experiments.