SCOREPLUS: Scalable Node Representation Learning in Customer Networks: New Techniques and Application. KU Leuven
This dissertation aims to bridge the gap between state-of-the-art representation learning and business applications, in particular fraud detection.
Representation learning is a structured and automated feature engineering approach, circumventing manual feature creation, which is inflexible, expensive, tedious, and error-prone. We focus on network representation learning techniques that transform a non-Euclidean network topology into a ...