Verifiable Noise-Based Privacy-Preserving Scheme for Trustworthy Consumer Applications

Many consumer applications are developed by collecting massive amounts of data from consumer devices. However, it also introduces complex challenges in safeguarding sensitive data, maintaining privacy, and defending against a wide spectrum of privacy leakage risks. Among the existing approaches, data perturbation is one of the most effective methods for protecting consumer data privacy. Nevertheless, two critical challenges remain unresolved: (1) how to ensure that the added noise strictly follows the required specifications; and (2) how to obtain accurate data. To address these issues, this paper proposes a Verifiable Noise-based Privacy Preserving (VNPP) scheme to improve data quality for trustworthy consumer applications. First, a Data Disturbance Verifiable (DDV) mechanism is designed to inject noise in a predetermined manner by employing blockchain technology. Second, a noise elimination approach is proposed to recover truthful data, thereby overcoming the limitation of traditional data perturbation methods in which noise cannot be effectively removed. Third, a data quality enhancement strategy is developed, which leverages interval verification to filter out low-quality data and select reliable users, thereby improving data quality under privacy protection. Extensive experiments demonstrate that the proposed VNPP scheme outperforms existing methods in terms of both robustness and accuracy. Its advantages lie in being more effective at eliminating noise, filtering malicious data, and implementing a more rational strategy for weighting trustworthy data. On average, VNPP achieves performance improvements of 86.29% in MAE and 86.90% in RMSE compared to the baseline algorithms.