[Paper Review] Online Pricing with Reserve Price Constraint for Personal Data Markets

Chaoyue Niu, Zhenzhe Zheng+3 more|arXiv (Cornell University)|Nov 28, 2019

Privacy-Preserving Technologies in Data56 references4 citations

TL;DR

This paper proposes a contextual dynamic pricing mechanism with reserve price constraint for online personal data markets, using ellipsoidal optimization to minimize regret in revenue maximization. It achieves logarithmic worst-case regret under low uncertainty and demonstrates effectiveness across data markets, hospitality, and advertising applications with low latency and memory overhead.

ABSTRACT

The society's insatiable appetites for personal data are driving the emergency of data markets, allowing data consumers to launch customized queries over the datasets collected by a data broker from data owners. In this paper, we study how the data broker can maximize her cumulative revenue by posting reasonable prices for sequential queries. We thus propose a contextual dynamic pricing mechanism with the reserve price constraint, which features the properties of ellipsoid for efficient online optimization, and can support linear and non-linear market value models with uncertainty. In particular, under low uncertainty, our pricing mechanism provides a worst-case regret logarithmic in the number of queries. We further extend to other similar application scenarios, including hospitality service, online advertising, and loan application, and extensively evaluate three pricing instances of noisy linear query, accommodation rental, and impression over MovieLens 20M dataset, Airbnb listings in U.S. major cities, and Avazu mobile ad click dataset, respectively. The analysis and evaluation results reveal that our proposed pricing mechanism incurs low practical regret, online latency, and memory overhead, and also demonstrate that the existence of reserve price can mitigate the cold-start problem in a posted price mechanism, and thus can reduce the cumulative regret.

Motivation & Objective

• To address the challenge of maximizing data broker revenue in online personal data markets with uncertain query values.
• To design a dynamic pricing mechanism that incorporates a reserve price constraint to mitigate the cold-start problem.
• To support both linear and non-linear market value models under uncertainty using efficient online optimization.
• To evaluate the mechanism across diverse real-world datasets in data markets, hospitality, and online advertising.
• To demonstrate low practical regret, online latency, and memory efficiency in dynamic pricing scenarios.

Proposed method

• Employs an ellipsoidal knowledge set to represent uncertainty in market value estimation, enabling efficient online optimization.
• Uses a contextual dynamic pricing framework where the data broker posts prices based on query context and historical feedback.
• Applies a reserve price constraint that sets a minimum threshold for pricing, preventing underpricing and reducing initial regret.
• Implements a gradient-based online learning algorithm with ellipsoid updates to maintain and refine the belief over the weight vector of market values.
• Extends the mechanism to non-linear market value models by transforming the feature space and maintaining uncertainty via ellipsoidal approximations.
• Introduces a feedback mechanism where only binary feedback (sale/no sale) is observed, enabling operation under partial information.

Experimental results

Research questions

• RQ1 How can a data broker maximize cumulative revenue in online personal data markets with uncertain query values?
• RQ2 What is the impact of incorporating a reserve price constraint on reducing cumulative regret in posted price mechanisms?
• RQ3 Can an ellipsoidal optimization framework achieve low regret in dynamic pricing with limited feedback and high-dimensional contexts?
• RQ4 How does the reserve price help mitigate the cold-start problem in online pricing systems?
• RQ5 To what extent does the proposed mechanism generalize across different application domains like data markets, hospitality, and online advertising?

Key findings

• The proposed mechanism achieves a worst-case regret that is logarithmic in the number of queries under low uncertainty, significantly outperforming non-reserve-based approaches.
• The inclusion of a reserve price constraint effectively mitigates the cold-start problem, reducing initial regret and accelerating convergence.
• Empirical evaluation on MovieLens 20M, Airbnb U.S. listings, and Avazu ad click datasets shows low practical regret, low online latency, and minimal memory overhead.
• The mechanism supports both linear and non-linear market value models while maintaining computational efficiency through ellipsoidal approximation.
• The feedback mechanism, based only on binary sale/no-sale outcomes, enables robust performance despite limited information.
• The results validate that the reserve price constraint enhances revenue optimization by preventing underpricing and improving early decision quality.