[Statlist] Research Seminar in Statistics | *FRIDAY 6 OCTOBER 2023* | GSEM, University of Geneva

gsem-support-instituts g@em-@upport-|n@t|tut@ @end|ng |rom un|ge@ch
Mon Oct 2 09:23:27 CEST 2023


Dear All,

We are pleased to invite you to our next Research Seminar, organized by Professor Sebastian Engelke on behalf of the Research Center for Statistics
< https://www.unige.ch/gsem/en/research/institutes/rcs/team/ >.

FRIDAY 6 OCTOBER 2023, at 11:15 am, Uni Mail M 5220 & Online
Zoom research webinar: https://unige.zoom.us/j/92924332087?pwd=U1U1NFk4dTFCRHBMeWYrSDBQcXBiQT09
Meeting ID: 929 2433 2087
Passcode: 399192

A Joint Model of Customers' Attrition, Transaction, and Spending Process in non-Contractual Business Settings
Markus MEIERER, GSEM, & Jeffrey NÄF, INRIA, France
< https://www.unige.ch/gsem/en/research/faculty/all/markus-meierer/ > et < https://ch.linkedin.com/in/jeffrey-naef >

ABSTRACT:
Modeling customer purchases is a complex task, particularly in retail and other non-contractual business settings. Often, the focus is on capturing three processes, i.e., customers' attrition, transaction, and spending behavior. The latter is generally modeled independently of the other two processes. In the literature, this approach has regularly been challenged. However, no joint modeling approach has yet been proposed that provides comparable flexibility, robustness, and accessibility to current practices. In this talk, we present a novel probabilistic model, that combines all 3 processes, while still admitting a closed-form likelihood. The new model can thus be estimated efficiently, is flexible, and even allows for the inclusion of time-invariant and time-varying covariates. While benchmarking with real-world data from the retailing industry is still ongoing, we give some insights into the improved performance of our model in predicting customer behavior.

> View the Research Seminar agenda: < https://www.unige.ch/gsem/en/research/seminars/rcs/ >

Regards,

Sandra Vuadens
Assistant to research and institutes
gsem-support-instituts using unige.ch



More information about the Statlist mailing list