A Method for Quantifying Data Utility in Privacy-Preserving Scenarios

Data masking is a critical component in the realm of data security, playing a significant role in preventing the involuntary disclosure of private information within data and ensuring that data is used legally. However, data masking achieves security by obfuscating the information contained in the original data, which can lead to a decrease in data usability while safeguarding it. To mitigate the negative impacts of data masking, we have devised a novel quantitative method for assessing the usability of masked data, termed QM-ADUGN (Quantitative Method for Masked Data Usability with Generalization and Noise). This method targets two common masking techniques, data generalization and noise addition, and quantifies data usability from two perspectives: accuracy and complexity. We have also designed experiments to validate the effectiveness of this usability quantification method. The experimental results indicate that the Pearson correlation coefficient between the quantification results and the actual usability of the data exceeds 0.8 in various scenarios, demonstrating a very strong correlation. Therefore, this quantification method can effectively reflect the true usability of the data.