Modern Data Pricing Models: Taxonomy and Comprehensive Survey

Xiaoye Miao,Huanhuan Peng,Xinyu Huang,Lu Chen,Yunjun Gao,Jianwei Yin

DOI: https://doi.org/10.48550/arXiv.2306.04945

2023-06-08

Databases

Abstract:Data play an increasingly important role in smart data analytics, which facilitate many data-driven applications. The goal of various data markets aims to alleviate the issue of isolated data islands, so as to benefit data circulation. The problem of data pricing is indispensable yet challenging in data trade. In this paper, we conduct a comprehensive survey on the modern data pricing solutions. We divide the data pricing solutions into three major strategies and thirteen models, including query pricing strategy, feature-based data pricing strategy, and pricing strategy in machine learning. It is so far the first attempt to classify so many existing data pricing models. Moreover, we not only elaborate the thirteen specific pricing models within each pricing strategy, but also make in-depth analyses among these models. We also conclude five research directions for the data pricing field, and put forward some novel and interesting data pricing topics. This paper aims at gaining better insights, and directing the future research towards practical and sophisticated pricing mechanisms for better data trade and share.

What problem does this paper attempt to address?

The paper aims to address the issue of data pricing in modern data markets. With the advent of the big data era, the role of data in intelligent data analysis has become increasingly important, but the phenomenon of data silos limits the effective flow of data. To promote data sharing, the emergence of modern data markets has become a bridge connecting data owners and data buyers. However, how to reasonably price data in the process of data transactions is an indispensable and challenging problem. Data pricing is not only the foundation of data market operations but also crucial for the formation of incentive mechanisms.

This paper proposes a comprehensive classification system for data pricing solutions, covering three main strategies and 13 specific models, including query pricing strategies, feature-based data pricing strategies, and pricing strategies in machine learning. This is the first attempt to classify such a diverse range of data pricing methods. In addition, the paper not only provides a detailed introduction to these 13 specific pricing models but also deeply analyzes the connections between these models and proposes five future research directions and some novel and interesting data pricing topics. Through this research, the authors hope to provide useful insights for the future development of the data pricing field and promote the study of practical and complex pricing mechanisms.

In summary, this paper attempts to address the issue of data pricing in modern data markets and proposes a new classification framework to cover various data pricing methods, with the aim of promoting the practical application of data trade and sharing.