Please note the time and location for each event in the schedule below.
For questions about the seminar schedule, please contact Antai Wang.
Date
Location
Speaker, Affiliation, and Title
Host
April 3
2:30PM / CULM LHI
Joshua Loftus, Stern School of Business, New York University Conditional Inference after Model Selection for Significance and Goodness of Fit Tests
We study conditional inference for error control of hypothesis tests conducted after model selection. Since selection methods choose the "best" model that fits the data, significance tests for variables in that model will tend to be anti-conservative, and goodness of fit tests will tend to be conservative. This is troubling, as it implies these tests in practice do not actually provide evidence in favor of the chosen variables or model. We demonstrate methods of post-selection inference to obtain conditionally valid significance tests and conditionally unbiased goodness of fit tests and show how these outperform unadjusted tests.
Wenge Guo
April 25
4:00PM / CULM LHI
Qingfeng Liu, Department of Economics, Otaru University of Commerce, Japan Model Averaging Estimation for Nonlinear Regression Models
This paper considers the problem of model averaging estimation for regression mod- els that can be nonlinear in their parameters and variables. We develop a nonlinear model averaging method (NMA) and propose the weight-choosing criterion NICMA. We show that NICMA is an asymptotically unbiased estimator of the risk function with an additional constant term under nonlinear settings. We also prove the optimality of NMA for candidate model sets with fixed or increasing number of candidate models, the convergence of the selected weight, and obtain the asymptotic distribution of the out-of-sample forecast. Simulation results reveal that NMA lead to relatively lower risks compared with alternative model selection and model averaging methods in most situations.