Differential privacy statistical inference for a directed graph network model with covariates

The real network has two characteristics: heterogeneity and homogeneity. A directed network model with covariates is proposed to analyze these two features, and the asymptotic theory of parameter Maximum likelihood estimators(MLEs) is established. However, in many practical cases, network data often carries a lot of sensitive information. How to achieve the trade-off between privacy and utility has become an important issue in network data analysis. In this paper, we study a directed $\beta$-model with covariates under differential privacy mechanism. It includes $2n$-dimensional node degree parameters $\boldsymbol{\theta}$ and a $p$-dimensional homogeneity parameter $\boldsymbol{\gamma}$ that describes the covariate effect. We use the discrete Laplace mechanism to release noise for the bi-degree sequences. Based on moment equations, we estimate the parameters of both degree heterogeneity and homogeneity in the model, and derive the consistency and asymptotic normality of the differentially private estimators as the number of nodes tends to infinity. Numerical simulations and case studies are provided to demonstrate the validity of our theoretical results.