Change-Point Detection in Dynamic Networks with Missing Links

TL;DR

This paper introduces a robust change-point detection method for dynamic networks with missing links, using a Matrix CUSUM test that is minimax optimal and scalable to large networks.

Contribution

It proposes a new change-point detection approach based on Matrix CUSUM for partially observed networks, with proven optimality and robustness to missing data.

Findings

The method is minimax optimal in detecting change-points.

It performs well in simulations and real data applications.

The approach is robust to missing links in large networks.

Abstract

Structural changes occur in dynamic networks quite frequently and its detection is an important question in many situations such as fraud detection or cybersecurity. Real-life networks are often incompletely observed due to individual non-response or network size. In the present paper we consider the problem of change-point detection at a temporal sequence of partially observed networks. The goal is to test whether there is a change in the network parameters. Our approach is based on the Matrix CUSUM test statistic and allows growing size of networks. We show that the proposed test is minimax optimal and robust to missing links. We also demonstrate the good behavior of our approach in practice through simulation study and a real-data application.