Abstract:Numerous imputation methods have been developed for missing data. However, these methods apply mainly to independent data, and the assumption of independence disregards connections of units through social relationships (e.g., friendship, follower relationship). In fact, observed responses from connected friends should provide valuable information for missing responses. This motivates us to conduct an imputation by borrowing information from connected friends using a network structure. With the missing assumption and using observed information only, we propose a partial likelihood approach and develop the corresponding maximum partial likelihood estimator (MPLE). The estimator's consistency and asymptotic normality are established. Using the MPLE, we then develop a novel regression imputation method. The method utilizes both auxiliary information and connected complete units (i.e., network information); using the imputed data, we can compute the sample mean of the responses. We show this method to be consistent and asymptotically normal. Compared with the imputation method using auxiliary information only (i.e., ignoring network information), the proposed estimator is statistically more efficient. Extensive simulation studies are conducted to demonstrate the finite-sample performance of the proposed method. We then analyze a real example about QQ in mainland China.

Network Autoregression for Incomplete Matrix-Valued Time Series

On a matrix‐valued autoregressive model

Network Imputation for a Spatial Autoregression Model with Incomplete Data

Estimation of Grouped Time-Varying Network Vector Autoregression Models

Banded Spatio-Temporal Autoregressions

Reduced-Rank Matrix Autoregressive Models: A Medium $N$ Approach

Graphical Assistant Grouped Network Autoregression Model: a Bayesian Nonparametric Recourse

A NAIVE LEAST SQUARES METHOD FOR SPATIAL AUTOREGRESSION WITH COVARIATES

Covariance Matrix Estimation Via Network Structure

Statistica Sinica Preprint No : SS-2017-0135 Title A Naive Least Squares Method for Spatial Autoregression with Covariates

Embedding Network Autoregression for time series analysis and causal peer effect inference

Two-mode Network Autoregressive Model for Large-Scale Networks

Two-way Threshold Matrix Autoregression

Dynamic Network Quantile Regression Model

Network Vector Autoregressive Model for Dyadic Response Variables

Feature Screening for Network Autoregression Model

A Latent Moving Average Model for Network Regression

Least Squares Inference for Data with Network Dependency

Modelling, Detrending and Decorrelation of Network Time Series

Semiparametric Modeling and Analysis for Longitudinal Network Data