Node classification for signed networks using diffuse interface methods

by   Jessica Bosch, et al.
TU Chemnitz
Universität Saarland
The University of British Columbia

Signed networks are a crucial tool when modeling friend and foe relationships. In contrast to classical undirected, weighted graphs, the edge weights for signed graphs are positive and negative. Crucial network properties are often derived from the study of the associated graph Laplacians. We here study several different signed network Laplacians with a focus on the task of classifying the nodes of the graph. We here extend a recently introduced technique based on a partial differential equation defined on the signed network, namely the Allen-Cahn-equation, to classify the nodes into two or more classes. We illustrate the performance of this approach on several real-world networks.


Optimization of a partial differential equation on a complex network

Differential equations on metric graphs can describe many phenomena in t...

A comparison of PINN approaches for drift-diffusion equations on metric graphs

In this paper we focus on comparing machine learning approaches for quan...

On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features

While Graph Neural Networks (GNNs) have recently become the de facto sta...

All Graphs with at most 8 nodes are 2-interval-PCGs

A graph G is a multi-interval PCG if there exist an edge weighted tree T...

Predicting Positive and Negative Links with Noisy Queries: Theory & Practice

Social networks and interactions in social media involve both positive a...

Weighted, Bipartite, or Directed Stream Graphs for the Modeling of Temporal Networks

We recently introduced a formalism for the modeling of temporal networks...

Network Classification in Temporal Networks Using Motifs

Network classification has a variety of applications, such as detecting ...

Please sign up or login with your details

Forgot password? Click here to reset