Personalized Dynamic Pricing with Machine Learning: High-Dimensional Features and Heterogeneous Elasticity

Ban, Gah‐Yi; Keskin, N. Bora

doi:10.1287/mnsc.2020.3680

Cited by 127 publications

(81 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When the agent makes decisions in the process of data collection, the datadriven approach is usually associated with a framework that integrates such a process with decision making, so that the agent is learning the unknown parameters or environment while maximizing revenues 1 . The previous works by Vayanos (2017), Zhang et al (2020), Cohen et al (2018, Ettl et al (2020), Ban and Keskin (2020) fall into this category. It is connected to a large body of literature on demand learning and dynamic pricing.…”

Section: Related Workmentioning

confidence: 95%

Model-Free Assortment Pricing with Transaction Data

Chen¹,

Ciré²,

Hu³

et al. 2021

Preprint

View full text Add to dashboard Cite

We study a problem in which a firm sets prices for products based on the transaction data, i.e., which product past customers chose from an assortment and what were the historical prices that they observed.Our approach does not impose a model on the distribution of the customers' valuations and only assumes, instead, that purchase choices satisfy incentive-compatible constraints. The individual valuation of each past customer can then be encoded as a polyhedral set, and our approach maximizes the worst-case revenue assuming that new customers' valuations are drawn from the empirical distribution implied by the collection of such polyhedra. We show that the optimal prices in this setting can be approximated at any arbitrary precision by solving a compact mixed-integer linear program. Moreover, we study special practical cases where the program can be solved efficiently, and design three approximation strategies that are of low computational complexity and interpretable. Comprehensive numerical studies based on synthetic and real data suggest that our pricing approach is uniquely beneficial when the historical data has a limited size or is susceptible to model misspecification.

show abstract

Section: Related Workmentioning

confidence: 95%

Model-Free Assortment Pricing with Transaction Data

Chen¹,

Ciré²,

Hu³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…and the competitor's rate as the feature vector for each customer. Note that a description of the data set (with descriptive statistics on the demand and available features) is available in Ban and Keskin (2020). The objective is to offer a personalized lending price (from a range of choices) based on personal information such as FICO score to a customer who will either accept or reject it.…”

Section: D2 Proof Of Lemma D3mentioning

confidence: 99%

“…We use On-Line Auto Lending dataset CRPM-12-001 in our real data case study 2 . We use the same features selection as in Ban and Keskin (2020); Cheung et al (2018) in the dataset and select FICO score, the term of contract, the loan amount approved, prime rate, the type of car,…”

Section: A Randomness Conditionmentioning

confidence: 99%

Generalized Linear Bandits with Local Differential Privacy

Han,

Liang,

Wang

et al. 2021

Preprint

View full text Add to dashboard Cite

Contextual bandit algorithms are useful in personalized online decision-making. However, many applications such as personalized medicine and online advertising require the utilization of individual-specific information for effective learning, while user's data should remain private from the server due to privacy concerns. This motivates the introduction of local differential privacy (LDP), a stringent notion in privacy, to contextual bandits. In this paper, we design LDP algorithms for stochastic generalized linear bandits to achieve the same regret bound as in non-privacy settings. Our main idea is to develop a stochastic gradient-based estimator and update mechanism to ensure LDP. We then exploit the flexibility of stochastic gradient descent (SGD), whose theoretical guarantee for bandit problems is rarely explored, in dealing with generalized linear bandits. We also develop an estimator and update mechanism based on Ordinary Least Square (OLS) for linear bandits. Finally, we conduct experiments with both simulation and real-world datasets to demonstrate the consistently superb performance of our algorithms under LDP constraints with reasonably small parameters (ε, δ) to ensure strong privacy protection.

show abstract

“…In particular, in the eld of revenue management, there have been many studies on demand learning and price experimentation using this framework, since the early seminal works by Araman and Caldentey (2009), Besbes and Zeevi (2009), Broder and Rusmevichientong (2012). Many extensions and new features have been studied, such as network revenue management (Besbes and Zeevi 2012, Ferreira et al 2018, personalized dynamic pricing (Chen and Gallego 2020, Miao et al 2019, Ban and Keskin 2020, limited price experimentations (Cheung et al 2017), and non-stationarity (Besbes et al 2015). Readers may refer to den Boer (2015) for a review of papers in this area.…”

Section: Online Learning and Multi-armed Banditmentioning

confidence: 99%

Revenue Maximization and Learning in Products Ranking

Chen

Yang

2020

Preprint

View full text Add to dashboard Cite

We consider the revenue maximization problem for an online retailer who plans to display a set of products di ering in their prices and qualities and rank them in order. The consumers have random attention spans and view the products sequentially before purchasing a "satis cing" product or leaving the platform emptyhanded when the attention span gets exhausted. Our framework extends the cascade model in two directions: the consumers have random attention spans instead of xed ones and the rm maximizes revenues instead of clicking probabilities. We show a nested structure of the optimal product ranking as a function of the attention span when the attention span is xed and design a 1/𝑒-approximation algorithm accordingly for the random attention spans. When the conditional purchase probabilities are not known and may depend on consumer and product features, we devise an online learning algorithm that achieves Õ ( √ 𝑇 ) regret relative to the approximation algorithm, despite of the censoring of information: the attention span of a customer who purchases an item is not observable. Numerical experiments demonstrate the outstanding performance of the approximation and online learning algorithms.

show abstract

Personalized Dynamic Pricing with Machine Learning: High-Dimensional Features and Heterogeneous Elasticity

Cited by 127 publications

References 18 publications

Model-Free Assortment Pricing with Transaction Data

Model-Free Assortment Pricing with Transaction Data

Generalized Linear Bandits with Local Differential Privacy

Revenue Maximization and Learning in Products Ranking

Contact Info

Product

Resources

About