Towards a Secure and Efficient Framework for Privacy-Preserving Truth Discovery in Sparse Data Scenario

Mobile crowd sensing (MCS) technology has demonstrated significant value across a wide range of practical applications by leveraging data collected from numerous mobile devices. However, the data gathered through such devices often contains inherent inaccuracies. To derive insights that closely approximate the ground truth from this potentially unreliable data, truth discovery techniques have been developed. At its core, truth discovery involves evaluating the credibility of participants and aggregating their data accordingly. Existing privacy-preserving truth discovery frameworks, however, fail to adequately address privacy concerns in these sparse data environments. In this paper, we systematically define the privacy issues that may arise during the truth discovery process in sparse data settings. To address these challenges, we propose a novel privacy-preserving truth discovery scheme tailored for sparse data scenarios. Through rigorous theoretical analysis and extensive experimental validation, we demonstrate that the proposed scheme not only meets stringent privacy protection requirements but also ensures low computational and communication overhead.