基于信息熵与风险厌恶视角的政府数据隐私计量与分级模型OA

To address the limitations of traditional privacy measurement models in government data privacy protection,specifi-cally their neglect of data subjects'subjective preferences and insufficient cross-scenario robustness,this paper proposes a pri-vacy measurement and classification model based on information entropy and risk aversion.The model aims to achieve precise quantification of privacy risks and enhance the adaptability of classification strategies across different domains.By constructing a multi-level privacy element classification system,the model quantifies the inherent privacy information of data using information entropy theory.It further introduces a risk aversion mechanism for privacy leakage and designs a dynamic weight allocation strat-egy to respond to the core privacy needs of different departments.The model integrates regularization modules,privacy element measurement modules,and risk aversion calibration modules to collaboratively compute comprehensive privacy risk values.Ex-perimental results based on multi-source heterogeneous synthetic datasets demonstrate that the model significantly outperforms existing methods in terms of privacy measurement accuracy,privacy element sensitivity,and cross-scenario adaptability.The findings indicate that the synergy between information entropy and dynamic risk aversion weights enables multi-dimensional pre-cise assessment of privacy risks,providing a scientific basis for differentiated privacy protection in government data sharing and openness.