SEMINARS
Surrogate Space Based Dimension Reduction for Nonignorable Nonresponse

2020-01-10　10:00 — 11:00

639

Sufficient dimension reduction (SDR) for nonignorable nonresponse  poses a challenge and up to now no literature is available on this problem. In the nonignorable case, the SDR methods developed for ignorable missing data generally yield serious estimation bias and thus are invalid. In this article, a regression-calibration-based cumulative mean estimation (RC-CUME) procedure is proposed to recover the central subspace $\mathcal S_{Y|\mathbf X}$ with the aid of a surrogate subspace. Asymptotic properties of the RC-CUME are investigated. A modified BIC-type criterion is used to determine the structural dimension of $\mathcal S_{Y|\mathbf X}$. We also extend our procedure to some other SDR methods. Simulation studies are conducted to access the finite-sample performance of the proposed RC-CUME approach, and a real data set is analyzed for illustration.