Instance-dependent cost-sensitive learning for detecting transfer fraud

Höppner, Sebastiaan; Baesens, Bart; Verbeke, Wouter; Verdonck, Tim

doi:10.1016/j.ejor.2021.05.028

Cited by 39 publications

(28 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A third stream concerns causal model learning. While the proposed framework allows factoring in costs and benefits in a post-processing stage, similar to cost-sensitive learning methods, they may be taken into account during causal model development (Höppner et al, 2022). Additionally, the effect of cost-sensitive causal learning on the performance and stability of causal models in cases of imbalanced data sets, both in terms of the class and the treatment distribution, remains to be explored.…”

Section: Discussionmentioning

confidence: 99%

To do or not to do? Cost-sensitive causal classification with individual treatment effect estimates

Verbeke¹,

Olaya²,

Guerry³

et al. 2023

European Journal of Operational Research

Self Cite

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

To do or not to do? Cost-sensitive causal classification with individual treatment effect estimates

Verbeke¹,

Olaya²,

Guerry³

et al. 2023

European Journal of Operational Research

Self Cite

View full text Add to dashboard Cite

“…Instead, it requires modellers to predefine (the number of) customer segments. This while instance-dependent cost sensitive learning has already proven to work in related fields such as fraud detection (Höppner et al, 2021). Therefore, we argue that a customer-specific optimization is better suited to the heterogeneous B2B customer base and is key to our proposed implementation.…”

Section: Customer Churn Prediction In Business-to-businessmentioning

confidence: 97%

B2Boost: instance-dependent profit-driven modelling of B2B churn

et al. 2022

View full text Add to dashboard Cite

The purpose of this paper is to enhance current practices in business-to-business (B2B) customer churn prediction modelling. Following the recent trend from accuracy-based to profit-driven evaluation businessto-customer churn prediction, we present a novel expected maximum profit measure for B2B customer churn (EMPB), which is used to demonstrate how current practices are suboptimal due to large discrepancies in customer value. To directly incorporate the heterogeneity of customer values and profit concerns of the company, we propose an instance-dependent profit maximizing classifier based on gradient boosting, named B2Boost. The main innovation of B2Boost is the fact that it considers these differences and incorporates them into the model construction by maximizing the objective function in terms of the EMPB. The results indicate that the expected maximal profit gains made in our analyses are substantial.This study arguments towards both deploying models based on customer-specific profitability differences, as well as evaluating based on our instance-dependent EMPB measure.

show abstract

“…The works regarding the proposal of new example-dependent cost-sensitive learning methods present comparison evaluations between their respective cost-insensitive versions of the methods [1,2,3,4,13,27]. These works use datasets from banks and credit card companies, for applications of fraud detection, credit scoring and direct marketing analysis.…”

Section: Related Workmentioning

confidence: 99%

“…Regarding the works that propose novel cost-sensitive methods, none of them applied the methods on a dataset originated from tax administration systems [1,2,3,4,13,27]. The experimental evaluation undertaken in this present work includes a set of methods with the best performances presented in each one of the referred related works.…”

Section: Related Workmentioning

confidence: 99%

On the evaluation of example-dependent cost-sensitive models for tax debts classification

Lima¹,

Fernandes²,

Moura³

2022

Anais Do XIX Encontro Nacional De Inteligência Artificial E Computacional (ENIAC 2022)

View full text Add to dashboard Cite

Example-dependent cost-sensitive classification methods are suitable to many real-world classification problems, where the costs, due to misclassification, vary among every example of a dataset. Tax administration applications are included in this segment of problems, since they deal with different values involved in the tax payments. To help matters, this work presents an experimental evaluation which aims to verify whether cost-sensitive learning algorithms are more cost-effective on average than traditional ones. This task is accomplished in a tax administration application domain, what implies the need of a cost-matrix regarding debt values. The obtained results show that cost-sensitive methods avoid situations like erroneously granting a request with a debt involving millions of reals. Considering the savings score, the cost-sensitive classification methods achieved higher results than their traditional method versions.

show abstract

Instance-dependent cost-sensitive learning for detecting transfer fraud

Cited by 39 publications

References 29 publications

To do or not to do? Cost-sensitive causal classification with individual treatment effect estimates

To do or not to do? Cost-sensitive causal classification with individual treatment effect estimates

B2Boost: instance-dependent profit-driven modelling of B2B churn

On the evaluation of example-dependent cost-sensitive models for tax debts classification

Contact Info

Product

Resources

About