Completing Networks by Learning Local Connection Patterns

TL;DR

This paper introduces C-GIN, a graph auto-encoder model that leverages local structural patterns to improve network completion, outperforming baseline models especially on networks with higher Reachable Clustering Coefficient.

Contribution

The paper presents a novel C-GIN model that captures local connection patterns for network completion, incorporating a new metric Reachable Clustering Coefficient for better performance assessment.

Findings

C-GIN achieves higher accuracy than baseline models in network completion.

The model performs better on networks with higher Reachable CC.

Less information is needed for effective network completion with C-GIN.

Abstract

Network completion is a harder problem than link prediction because it does not only try to infer missing links but also nodes. Different methods have been proposed to solve this problem, but few of them employed structural information - the similarity of local connection patterns. In this paper, we propose a model named C-GIN to capture the local structural patterns from the observed part of a network based on the Graph Auto-Encoder framework equipped with Graph Isomorphism Network model and generalize these patterns to complete the whole graph. Experiments and analysis on synthetic and real-world networks from different domains show that competitive performance can be achieved by C-GIN with less information being needed, and higher accuracy compared with baseline prediction models in most cases can be obtained. We further proposed a metric "Reachable Clustering Coefficient(CC)" based on network structure. And experiments show that our model perform better on a network with higher Reachable CC.