Uncertainty Quantification for Demand Prediction in Contextual Dynamic Pricing

Data-driven sequential decision has found a wide range of applications in modern operations management, such as dynamic pricing, inventory control, and assortment optimization. Most existing research on data-driven focuses designing an online policy to maximize the revenue. However, uncertainty quantification underlying true model function (e.g., demand function), critical problem for practitioners, not been well explored. In this paper, using prediction pricing motivating example, we study constructing accurate confidence intervals function. The main challenge is that sequentially collected data leads significant distributional bias maximum likelihood estimator or empirical risk minimization estimate, making classical statistics approaches Wald's test no longer valid. We address by developing debiased approach provide asymptotic normality guarantee estimator. Based estimator, both point-wise uniform