International Journals of Multidisciplinary
Research Academy (IJMRA)

Abstract:
This paper presents a production-grade, real-time optimization framework for ad serving in retailer media platforms, integrating three complementary machine learning systems: (i) a gradient-boosted bid optimization model, (ii) deep neural networks for click-through rate (CTR) and conversion rate (CVR) prediction, and (iii) a Bayesian multi-armed bandit for adaptive creative selection.
Download full Lenght Paper......