Personalized Dynamic Pricing with Machine Learning

Ban, Gah‐Yi; Keskin, N. Bora

doi:10.2139/ssrn.2972985

Cited by 37 publications

(43 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We now turn to the on-line auto lending dataset. This dataset was first studied by Phillips et al (2015), and subsequently used to evaluate dynamic pricing algorithms by Ban and Keskin (2017).…”

Section: Real Data On Online Auto-lendingmentioning

confidence: 99%

See 1 more Smart Citation

Meta Dynamic Pricing: Learning Across Experiments

2019

View full text Add to dashboard Cite

Section: Real Data On Online Auto-lendingmentioning

confidence: 99%

“…Setup: Following the approach of Phillips et al (2015) and Ban and Keskin (2017), we impute the price of a loan as the net present value of future payments (a function of the monthly payment, customer rate, and term approved; we refer the reader to the cited references for details). The allowable price range in our experiment is [0, 300].…”

Section: Featuresmentioning

confidence: 99%

Meta Dynamic Pricing: Learning Across Experiments

2019

View full text Add to dashboard Cite

“…Personalized dynamic pricing can be regarded as a special case of learning with contextual information (Qiang and Bayati, 2016;Javanmard and Nazerzadeh, 2016;Cohen et al, 2016;Ban and Keskin, 2017;Chen and Gallego, 2018). The main difference of our paper from this stream of literature is summarized below.…”

Section: Literature Reviewmentioning

confidence: 99%

A Primal-dual Learning Algorithm for Personalized Dynamic Pricing with an Inventory Constraint

Chen

Gallego²

2018

SSRN Journal

View full text Add to dashboard Cite

A firm is selling a product to different types (based on the features such as education backgrounds, ages, etc.) of customers over a finite season with non-replenishable initial inventory. The type label of an arriving customer can be observed but the demand function associated with each type is initially unknown. The firm sets personalized prices dynamically for each type and attempts to maximize the revenue over the season. We provide a learning algorithm that is near-optimal when the demand and capacity scale in proportion. The algorithm utilizes the primal-dual formulation of the problem and learns the dual optimal solution explicitly. It allows the algorithm to overcome the curse of dimensionality (the rate of regret is independent of the number of types) and sheds light on novel algorithmic designs for learning problems with resource constraints.

show abstract

“…More recently, several papers investigate the pricing problem with unknown demand in the presence of covariates (Nambiar et al, 2016;Qiang and Bayati, 2016;Javanmard and Nazerzadeh, 2016;Cohen et al, 2016;Ban and Keskin, 2017 Ban and Keskin (2017), the near-optimal policy achieves regret O(s √ T ). In their parametric frameworks, the dependence of the regret on d is rather mild-it does not appear on the exponent of T .…”

Section: Literature Reviewmentioning

confidence: 99%

“…Without covariates, it has been shown that parametric and nonparametric methods can achieve the same rate of regret O( √ T ) (e.g., compare Besbes and Zeevi 2009;Wang et al 2014 andZeevi 2014;den Boer and Zwart 2014). Assuming a linear form of the covariate, Qiang and Bayati (2016); Javanmard and Nazerzadeh (2016); Ban and Keskin (2017) have shown that the best achievable rate of regret is still O( √ T ) or O(log(T )), depending on speci c assumptions. Our result demonstrates that lack of knowledge of the reward function's parametric form is extremely costly when the covariate is highdimensional.…”

Section: Introductionmentioning

confidence: 99%