Date of Award

5-2013

Document Type

Thesis

Degree Name

Master of Science (MS)

Legacy Department

Mathematical Science

Advisor

Luo, June

Committee Member

Gallagher , Colin

Committee Member

Gerard , Patrick

Abstract

Approximate factor models are popular in finance and economics. A key to effectively utilizing such a model is to accurately estimate the error covariance matrix. Errors related to certain predictors are expected to be correlated and this must be modeled effectively. Adaptive thresholding is a method for estimating the error covariance matrix of such a model. This method is described in detail and a simulation study sheds light on the behavior of this method under different sample sizes and parameterizations.

Share

COinS