Tree-based Cost-Sensitive Methods for Fraud Detection in Imbalanced Data - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

Tree-based Cost-Sensitive Methods for Fraud Detection in Imbalanced Data

Guillaume Metzler
Xavier Badiche
  • Fonction : Auteur
  • PersonId : 1037464
Brahim Belkasmi
  • Fonction : Auteur
  • PersonId : 1037465
Elisa Fromont
Amaury Habrard
Marc Sebban

Résumé

Bank fraud detection is a difficult classification problem where the number of frauds is much smaller than the number of genuine transactions. In this paper, we present cost sensitive tree-based learning strategies applied in this context of highly imbalanced data. We first propose a cost sensitive splitting criterion for decision trees that takes into account the cost of each transaction and we extend it with a decision rule for classification with tree ensembles. We then propose a new cost-sensitive loss for gradient boosting. Both methods have been shown to be particularly relevant in the context of imbalanced data. Experiments on a proprietary dataset of bank fraud detection in retail transactions show that our cost sensitive algorithms allow to increase the retailer's benefits by 1,43% compared to non cost-sensitive ones and that the gradient boosting approach outperforms all its competitors.
Fichier principal
Vignette du fichier
CSTree.pdf (359.29 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01895967 , version 1 (09-11-2018)

Identifiants

Citer

Guillaume Metzler, Xavier Badiche, Brahim Belkasmi, Elisa Fromont, Amaury Habrard, et al.. Tree-based Cost-Sensitive Methods for Fraud Detection in Imbalanced Data. IDA 2018 - 17th International Symposium on Intelligent Data Analysis, Oct 2018, ‘s-Hertogenbosch, Netherlands. pp.213-224, ⟨10.1007/978-3-030-01768-2_18⟩. ⟨hal-01895967⟩
206 Consultations
676 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More