|
Activity Number:
|
167
- Data Mining and Econometrics
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 10, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
|
Abstract #318098
|
|
|
Title:
|
Statistical Inference for Noisy Matrix Completion Incorporating Auxiliary Information
|
|
Author(s):
|
Shujie Ma and Yinchu Zhu and PoYao Niu*
|
|
Companies:
|
University of California, Riverside and Economics, Brandeis University and Department of Statistics, University of California, Riverside
|
|
Keywords:
|
matrix completion ;
low-rank decomposition;
latent factors;
auxiliary covariates;
simultaneous testing
|
|
Abstract:
|
This paper investigates statistical inference for noisy matrix completion in a semi-supervised model when auxiliary covariates are available. The model consists of two parts. One part is a low-rank matrix induced by unobserved latent factors; the other part models the effects of the observed covariates through a coefficient matrix which is composed of high-dimensional column vectors. We propose an iterative least squares (LS) estimation approach that fully enjoys a low computational cost. We show that we only need to iterate the LS estimation a few times, and the resulting entry-wise estimators of the target matrix and the coefficient matrix are guaranteed to have asymptotic normal distributions. As a result, a pointwise confidence interval and individual inference for each entry of the unknown matrices can be conducted. Moreover, we propose a simultaneous testing procedure with multiplier bootstrap for the high-dimensional coefficient matrix. This simultaneous inferential tool can help us further investigate the effects of auxiliary covariates for the prediction of all missing entries.
|
Authors who are presenting talks have a * after their name.
