Statistics Colloquium
September 30 (Fri) 11 a.m.
Alavi Commons Room, 6625 Everett Tower

Efficient estimation by combining large-dimensional moment conditions

Ryan Cho
Department of Statistics
Western Michigan University

Abstract: The generalized method of moments (GMM) is widely applicable when the likelihood function is difficult to specify, while moment conditions are easy to formulate. The GMM is powerful as it optimally combines valid moment conditions, yet it could perform poorly if there are too many moments relative to the sample size due to limitation in finite samples. We are motivated by the problem where the dimension of moments far exceeds the sample size. In this project, we have proposed a new objective function which selects an optimal number of linear combinations of moments. This allows one to reduce the dimensionality of available moments while retaining most of the information from data. Theoretical results have shown that an optimal number of principal components can be consistently selected without loss of efficiency. The proposed method can also be applied to estimate the inverse of the covariance matrix in high-dimensional data frameworks.

Bio: Ryan Cho is a faculty member of the Department of Statistics at Western Michigan University. He earned his Ph.D. degree from the University of Illinois in Urbana-Champaign.

All statistics graduate students are expected to attend.


Past colloquiums


Department of Statistics
3304 Everett Tower
Western Michigan University
Kalamazoo MI 49008-5152 USA
(269) 387-1420 | (269) 387-1419 Fax