Abstract
Job offers reveal employer preferences about capabilities required for future cyberdefense. We model such job openings as edges of a bipartite network of organizations and technologies. We propose and train a parsimonious prediction algorithm with extant job offer data to predict which capabilities firms will require up to six months from now. We compare the efficiency of our method across several unsupervised learning similarity-based algorithms and a supervised learning method to optimize model dynamics.
Research Paper
Closed Access research paper available here: Anticipating Cyberdefense Capability Requirements by Link Prediction Analysis.
BibTex
@Inbook{Moreno2023,
author="Moreno, Santiago Anton
and Percia David, Dimitri
and Mermoud, Alain
and Maillart, Thomas
and Mezzetti, Anita",
editor="Keupp, Marcus Matthias",
title="Anticipating Cyberdefense Capability Requirements by Link Prediction Analysis",
bookTitle="Cyberdefense: The Next Generation",
year="2023",
publisher="Springer International Publishing",
address="Cham",
pages="135--145",
abstract="Job offers reveal employer preferences about capabilities required for future cyberdefense. We model such job openings as edges of a bipartite network of organizations and technologies. We propose and train a parsimonious prediction algorithm with extant job offer data to predict which capabilities firms will require up to six months from now. We compare the efficiency of our method across several unsupervised learning similarity-based algorithms and a supervised learning method to optimize model dynamics.",
isbn="978-3-031-30191-9",
doi="10.1007/978-3-031-30191-9_9",
url="https://doi.org/10.1007/978-3-031-30191-9_9"
}