Differentially private feature selection

The privacy-preserving data analysis has been gained significant interest across several research communities. The current researches mainly focus on privacy-preserving classification and regression. However, feature selection is also an essential component for data analysis, which can be used to reduce the data dimensionality and can be utilized to discover knowledge, such as inherent variables in data. In this paper, in order to efficiently mine sensitive data, a privacy preserving feature selection algorithm is proposed and analyzed in theory based on local learning and differential privacy. We also conduct some experiments on benchmark data sets. The Experimental results show that our algorithm can preserve the data privacy to some extent.

I. Introduction

Privacy is the sensitive information that data owner reluctant to disclose, which has been a growing concern in medical records, financial records, web search histories and social network data. Thus, an emerging challenge for machine learning and data mining is how to learn from these data sets without privacy leak. The current researches mainly focus on privacy-preserving classification and regression [1]. However, feature selection is also one of the key problems in machine learning and data mining [2], [3]. Feature selection brings the immediate effects of speeding up a machine learning or data mining algorithm, improving learning accuracy, and enhancing model comprehensibility. Various studies show that features can be removed without performance deterioration [4]. More-over, feature selection also leads to better data visualization, reduction of measurement and storage requirements. So the feature selection with privacy preserving is a very important issue and need to be deeply addressed. Before introducing the concrete privacy preserving feature selection algorithm, we will simply discuss the common properties of feature selection and basic knowledge for privacy preserving.