Instance-level explanations for fraud detection: a case study

Research output: Contribution to conferencePaperAcademic

25 Downloads (Pure)

Abstract

Fraud detection is a difficult problem that can benefit from predictive modeling. However, the verification of a prediction is challenging; for a single insurance policy, the model only provides a prediction score. We present a case study where we reflect on different instance-level model explanation techniques to aid a fraud detection team in their work. To this end, we designed two novel dashboards combining various state-of-the-art explanation techniques. These enable the domain expert to analyze and understand predictions, dramatically speeding up the process of filtering potential fraud cases. Finally, we discuss the lessons learned and outline open research issues.
Original languageEnglish
Pages28-33
Number of pages6
Publication statusPublished - 19 Jun 2018
Event2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018) - Stockholmsmässan, Stockholm, Sweden
Duration: 14 Jul 201814 Oct 2018
Conference number: 3
https://sites.google.com/view/whi2018

Workshop

Workshop2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018)
Abbreviated titleWHI 2018
CountrySweden
CityStockholm
Period14/07/1814/10/18
OtherPart of W17 of IJCAI-ECAI 2018
Internet address

Fingerprint

Insurance

Keywords

  • Interpretability
  • Explanation
  • Machine learning
  • Sensitivity analysis
  • Local rule extraction
  • Instance-level explanations
  • Fraud detection
  • Case study

Cite this

Collaris, D. A. C., Vink, L. M., & van Wijk, J. J. (2018). Instance-level explanations for fraud detection: a case study. 28-33. Paper presented at 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018), Stockholm, Sweden.
Collaris, D.A.C. ; Vink, Leo M. ; van Wijk, J.J. / Instance-level explanations for fraud detection: a case study. Paper presented at 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018), Stockholm, Sweden.6 p.
@conference{35e3dbbcdace496c9de10ba20024fe73,
title = "Instance-level explanations for fraud detection: a case study",
abstract = "Fraud detection is a difficult problem that can benefit from predictive modeling. However, the verification of a prediction is challenging; for a single insurance policy, the model only provides a prediction score. We present a case study where we reflect on different instance-level model explanation techniques to aid a fraud detection team in their work. To this end, we designed two novel dashboards combining various state-of-the-art explanation techniques. These enable the domain expert to analyze and understand predictions, dramatically speeding up the process of filtering potential fraud cases. Finally, we discuss the lessons learned and outline open research issues.",
keywords = "Interpretability, Explanation, Machine learning, Sensitivity analysis, Local rule extraction, Instance-level explanations, Fraud detection, Case study",
author = "D.A.C. Collaris and Vink, {Leo M.} and {van Wijk}, J.J.",
year = "2018",
month = "6",
day = "19",
language = "English",
pages = "28--33",
note = "2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018), WHI 2018 ; Conference date: 14-07-2018 Through 14-10-2018",
url = "https://sites.google.com/view/whi2018",

}

Collaris, DAC, Vink, LM & van Wijk, JJ 2018, 'Instance-level explanations for fraud detection: a case study' Paper presented at 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018), Stockholm, Sweden, 14/07/18 - 14/10/18, pp. 28-33.

Instance-level explanations for fraud detection: a case study. / Collaris, D.A.C.; Vink, Leo M.; van Wijk, J.J.

2018. 28-33 Paper presented at 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018), Stockholm, Sweden.

Research output: Contribution to conferencePaperAcademic

TY - CONF

T1 - Instance-level explanations for fraud detection: a case study

AU - Collaris, D.A.C.

AU - Vink, Leo M.

AU - van Wijk, J.J.

PY - 2018/6/19

Y1 - 2018/6/19

N2 - Fraud detection is a difficult problem that can benefit from predictive modeling. However, the verification of a prediction is challenging; for a single insurance policy, the model only provides a prediction score. We present a case study where we reflect on different instance-level model explanation techniques to aid a fraud detection team in their work. To this end, we designed two novel dashboards combining various state-of-the-art explanation techniques. These enable the domain expert to analyze and understand predictions, dramatically speeding up the process of filtering potential fraud cases. Finally, we discuss the lessons learned and outline open research issues.

AB - Fraud detection is a difficult problem that can benefit from predictive modeling. However, the verification of a prediction is challenging; for a single insurance policy, the model only provides a prediction score. We present a case study where we reflect on different instance-level model explanation techniques to aid a fraud detection team in their work. To this end, we designed two novel dashboards combining various state-of-the-art explanation techniques. These enable the domain expert to analyze and understand predictions, dramatically speeding up the process of filtering potential fraud cases. Finally, we discuss the lessons learned and outline open research issues.

KW - Interpretability

KW - Explanation

KW - Machine learning

KW - Sensitivity analysis

KW - Local rule extraction

KW - Instance-level explanations

KW - Fraud detection

KW - Case study

M3 - Paper

SP - 28

EP - 33

ER -

Collaris DAC, Vink LM, van Wijk JJ. Instance-level explanations for fraud detection: a case study. 2018. Paper presented at 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018), Stockholm, Sweden.