A1527
Title: A finite-horizon mixture cure model with application to online flea market data
Authors: Yuji Komiyama - Tohoku University (Japan) [presenting]
Yasumasa Matsuda - Tohoku University (Japan)
Masakazu Ishihara - New York University (United States)
Abstract: A mixture cure model is proposed that latently divides a population based on event occurrence within a finite time horizon. Conventional mixture cure models rely on event occurrence over an infinite horizon, introducing untestable assumptions that often lead to issues with identifiability and interpretability. By shifting the estimand to a specific period of interest, the proposed approach reduces reliance on these infinite-tail assumptions and aligns interpretations more closely with finite-horizon decision-making objectives. Through simulation studies, the statistical properties of the proposed estimator are first evaluated, including estimation bias and variance. It is further shown that relying on conventional infinite-horizon models for finite-horizon decision-making can lead to erroneous judgments. Finally, the model is applied to transaction data from Mercari, a Japanese online flea market platform. The empirical results reveal that the proposed model identifies different significant variables compared to the conventional model, offering interpretations that better reflect seasonal variation in user behavior. These findings suggest that the finite-horizon framing is broadly applicable to settings where decision-makers operate within well-defined time windows, such as marketing campaigns or contract renewals, providing practitioners with a statistically rigorous yet interpretable tool for real-world decision-making.
