ENSURING DATA TRUTHFULNESS AND PRIVACY PRESERVATION IN DATA MARKET

Abstract

As a significant business paradigm, 
many online information platforms have 
emerged to satisfy society’s needs for person- 
specific data, where a service provider 
collects raw data from data contributors, and 
then offers value-added data services to data 
consumers. However, in the data trading 
layer, the data consumers face a pressing 
problem, i.e., how to verify whether the 
service provider has truthfully collected and 
processed data? Furthermore, the data 
contributors are usually unwilling to reveal 
their sensitive personal data and real identities 
to the data consumers. In this paper, we 
propose TPDM, which efficiently integrates 
Truthfulness and Privacy preservation in Data 
Markets. TPDM is structured internally in an 
Encrypt-then-Sign fashion, using partially 
homomorphic encryption and identity-based 
signature. It simultaneously facilitates batch 
verification, data processing, and outcome 
verification, while maintaining identity 
preservation and data confidentiality. We also 
instantiate TPDM with a profile matching 
service and a data distribution service, and 
extensively evaluate their performances on 
Yahoo! Music ratings dataset and 2009 RECS 
dataset, respectively. Our analysis and 
evaluation results reveal that TPDM achieves 
several desirable properties, while incurring 
low 
computation 
and 
communication 
overheads when supporting large-scale data 
markets.