Bayesian Inference of Differentially Private Datasets in Linear Regression Models

• Author / Creator

  Jiang, Yangdi

• The demand for large dataset and demand of privacy protection are in constantly conflicts as the balance between the two is hard to keep. Differential privacy is a mathematical rigor definition that provides the balance bewteen these two opposite sides. It's developed with the purpose of making privacy-preserving analysis/inference. Ever since the introduction of differential privacy, the literatures around it have been flourishing. However, the methodologies of statistical analysis and inference given the differentially private dataset are not much studied. In this thesis, we will tackle this problem using Bayesian method from the perspective of measurement error problems. Our simulation study shows that it outperforms the existing method (SIMEX) when applying to a similar specification of problem. Additionally, we will investigate briefly the question whether it's beneficial to generate multiple DP datasets.

• Subjects / Keywords

  • Bayesian Inference
  • Differential Privacy
  • Regression
  • Data Synthesis

• Graduation date

  Fall 2021

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-jmq9-m038

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.