Data-Driven Smooth Testsased on Karhunen-Loeve Expansion
Date Issued
2009
Date
2009
Author(s)
Lin, Tzu-Chi
Abstract
New data-driven smooth tests are proposed in this thesis. The new testsre proposed to eschew the downward weighting problem of the traditionalmnibus tests, and the new tests are constructed based on the componentsf Karhunen-Lo′eve expansion of limiting process. As examples, we constructests for the null hypothesis of stationarity, coefficient stability, symmetricynamics of quantile autoregressive model, and bivariate independence.imulation results show that, new tests have moderate size control and niceower performance for a wide range of alternatives. In contrast to traditionalmnibus tests, new tests are more robust to complex models and perform wellnder high-frequency alternatives.
Subjects
Cramer-von Mises test
Karhunen-Loeve Expansion
Neyman smooth test
orthonormal polynomial
integral equation
stationarity
structural change
quantile autoregressive
bivariate independence
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-98-R96323032-1.pdf
Size
23.53 KB
Format
Adobe PDF
Checksum
(MD5):258f21f74259b5758c2d45d0bfe0f5f2