Predicting New Customers' Risk Type in the Credit Card Market

Zhao, Yizhi; Ying, Zhao; Song, Inseong

doi:10.1509/jmkr.46.4.506

Cited by 31 publications

(18 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This finding holds for models estimated on panel data as well as models estimated on cross-sectional data. Prior work in a different setting has empirically established similar results (e.g., Zhao, Zhao, and Song 2009). We attribute these findings to the following: As we have observed, accounting for dynamics only is not sufficient;…”

Section: Discussionsupporting

confidence: 76%

No future without the past? Predicting churn in the face of customer privacy

Holtrop

Wieringa²,

Gijsenberg³

et al. 2017

International Journal of Research in Marketing

View full text Add to dashboard Cite

For customer-centric firms, churn prediction plays a central role in churn management programs.Methodological advances have emphasized the use of customer panel data to model the dynamic evolution of a customer base to improve churn predictions. However, pressure from policy makers and the public geared to reducing the storage of customer data has led to firms' ‗self-policing' by limiting data storage, rendering panel data methods infeasible. We remedy these problems by developing a method that captures the dynamic evolution of a customer base without relying on the availability past data. Instead, using a recursively updated model our approach requires only knowledge of past model parameters. This generalized mixture of Kalman filters model maintains the accuracy of churn predictions compared to existing panel data methods when data from the past is available. In the absence of past data, applications in the insurance and telecommunications industry establish superior predictive performance compared to simpler benchmarks. These improvements arise because the proposed method captures the same dynamics and unobserved heterogeneity present in customer databases as advanced methods, while achieving privacy preserving data minimization and data anonymization. We therefore conclude that privacy preservation does not have to come at the cost of analytical operations.

show abstract

Section: Discussionsupporting

confidence: 76%

No future without the past? Predicting churn in the face of customer privacy

Holtrop

Wieringa²,

Gijsenberg³

et al. 2017

International Journal of Research in Marketing

View full text Add to dashboard Cite

show abstract

“…The delinquency rates of consumers in the credit card market are consistently higher than those in the other parts of the loan market [4]. It is important for a credit card company to identify consumer segments (high risk or low risk) to develop targeted marketing strategies.…”

Section: Discussionmentioning

confidence: 99%

“…Though customers with a high risk of default are included in this segment, the most profits of credit card also come from them [4]. Revolvers contain a kind of customers who have delinquent records.…”

Section: Introductionmentioning

confidence: 99%

“…Moreover, this kind of cardholders can be further roughly divided into two groups. For some users, the ones who do not meet minimum payment requirement for a long time (more than three successive months) are defined as high-risk group [4,5]; for the other users, they just delinquent occasionally because the factors like they forget to repay and will eventually pay off their debts, which can be defined as low-risk group [6,20].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Customers’ Risk Type Prediction Based on Analog Complexing

Zhu

Zhang

et al. 2015

Procedia Computer Science

View full text Add to dashboard Cite

Credit card holder's behaviour may change over time, which would lead to the change of risk type. This paper introduces a new method, AnalogComplexing (AC), to predict consumer's risk type. Furthermore, the new method uses the observed historical process itself for prediction which does not need any information of input variables in unknown prediction period. The authors applied the new proposed method AC to a bank customer dataset from one city of Western China, and the empirical study shows that AC is significantly better than widely-used neural network in terms of prediction accuracy. The empirical results in indicates that AC is an effective method, and provide a new way to predict consumers' risk type.

show abstract

“…For example, Zhao et al (2009) Incorporating this thinking, we set up our model in the following section.…”

Section: Literature Reviewmentioning

confidence: 99%