Multi-Party Data Pricing for Complex Data Trading Markets: A Rubinstein Bargaining Approach

TL;DR

The paper tackles pricing in complex data trading markets with multiple sellers and buyers, where data monopolies and silos distort value. It introduces a multi-factor Rubinstein bargaining framework that jointly models buyer data utility, data quality, and a buyer dataset satisfaction metric to derive prices under incomplete information. Key contributions include formalizing buyer utility via Shapley values, a four-dimension data quality evaluation, and an analytically derived equilibrium price that accounts for platform mediation and data quality. The approach is validated through simulations on the 20 Newsgroups dataset, showing how monopolies, data quality, and platform transparency influence prices and profits, with practical implications for fair, demand-driven data markets.

Abstract

With the rapid development of Internet of Things (IoT) and artificial intelligence technologies, data has become an important strategic resource in the new era. However, the growing demand for data has exacerbated the issue of \textit{data silos}. Existing data pricing models primarily focus on single factors such as data quality or market demand, failing to adequately address issues such as data seller monopolies and the diverse needs of buyers, resulting in biased pricing that cannot meet the complexities of evolving transaction scenarios. To address these problems, this paper proposes a multi-party data pricing model based on the Rubinstein bargaining model. The model introduces buyer data utility indicators and data quality assessments, comprehensively considering factors such as the utility, accuracy, and timeliness of data sets, to more accurately evaluate their value to buyers. To overcome the limitations of single-factor models, this paper innovatively introduces the buyer data set satisfaction indicator, which reflects the overall satisfaction of buyers with data sets by integrating both data utility and quality assessments. Based on this, the model uses the Rubinstein bargaining model to simulate the pricing process between multiple sellers and multiple buyers, yielding pricing results that better align with market demands. Experimental results show that the proposed model effectively addresses the pricing imbalance caused by data monopolies and demonstrates good applicability and accuracy in multi-seller, multi-buyer transaction environments. This research provides an effective pricing mechanism for complex data trading markets and has significant theoretical and practical value in solving pricing issues in actual data transactions.

Figures (8)

• Figure 1: System framework.
• Figure 2: Hierarchy Structure of Buyer Data Set Satisfaction Indicators
• Figure 3: Three-Stage Bargaining Process Diagram.
• Figure 4: Logistic decay function.
• Figure 5: Data Quality Score and Multi-Party Additional Earnings Curve.