Differentially Private k-Nearest Neighbor Missing Data Imputation

Chris Clifton

doi:10.1145/3507952

Differentially Private k-Nearest Neighbor Missing Data Imputation

2022-04-09

4

PRE

AI

DOI：10.1145/3507952

Original request for help

Save

Abstract

En 中文

Using techniques employing smooth sensitivity, we develop a method for k-nearest neighbor missing data imputation with differential privacy. This requires bounding the number of data incomplete tuples that can have their data complete donor changed by making a single addition or deletion to the dataset. The multiplicity of a single individual's impact on an imputed dataset necessarily means our mechanisms require the addition of more noise than mechanisms that ignore missing data, but we show empirically that this is significantly outweighed by the bias reduction from imputing missing data.

Keywords:

Differential privacy

statistical disclosure limitation

private data cleaning

smooth sensitivity

Differentially Private k-Nearest Neighbor Missing Data Imputation

Abstract

Journal

Organization