Graph link prediction in computer networks using Poisson matrix factorisation

DOI10.1214/21-AOAS1540MaRDI QIDQ2170383zbMATH OpenOpenAlexWikidataFDO

Authors Francesco Sanna Passino, Melissa J. M. Turcotte, Nicholas A. Heard

Publication date 5 September 2022

Published in The Annals of Applied Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2001.09456

zbMATH Keywords

anomaly detectiondynamic networksnew link predictionPoisson matrix factorisationstatistical cybersecurityvariational inference

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Applications of statistics (62P99) Social networks; opinion dynamics (91D30)

Abstract: Graph link prediction is an important task in cyber-security: relationships between entities within a computer network, such as users interacting with computers, or system libraries and the corresponding processes that use them, can provide key insights into adversary behaviour. Poisson matrix factorisation (PMF) is a popular model for link prediction in large networks, particularly useful for its scalability. In this article, PMF is extended to include scenarios that are commonly encountered in cyber-security applications. Specifically, an extension is proposed to explicitly handle binary adjacency matrices and include known categorical covariates associated with the graph nodes. A seasonal PMF model is also presented to handle seasonal networks. To allow the methods to scale to large graphs, variational methods are discussed for performing fast inference. The results show an improved performance over the standard PMF model and other statistical network models.

Recommendations

• Link prediction in dynamic networks using random dot product graphs
• Link prediction via sparse Gaussian graphical model
• Link prediction in temporal networks: integrating survival analysis and game theory
• Graph regularized nonnegative matrix factorization for temporal link prediction in dynamic networks
• A network structural approach to the link prediction problem

Cites work

• A review of dynamic network models with latent variables
• A survey of statistical network models
• Bayesian latent variable models for mixed discrete outcomes
• Bilinear Mixed-Effects Models for Dyadic Data
• Content‐boosted matrix factorization techniques for recommender systems
• Flexible low-rank statistical modeling with missing data and side information
• Latent Space Approaches to Social Network Analysis
• Latent space models for dynamic networks
• Modeling item-item similarities for personalized recommendations on Yahoo! front page
• On Bayesian new edge prediction and anomaly detection in computer networks
• On the identifiability of Bayesian factor analytic models
• Pattern recognition and machine learning.
• Smooth neighborhood recommender systems
• Statistical inference on random dot product graphs: a survey
• Statistical methods for network surveillance
• Variational Bayesian inference for the latent position cluster model for network data

Cited in

(2)

• Learning social networks from text data using covariate information
• Changepoint detection on a graph of time series

This page was built for publication: Graph link prediction in computer networks using Poisson matrix factorisation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2170383)