Doctoral Dissertation
Common Features in Vector
Nonlinear Time Series Models
DAO LI
Statistics
Örebro Studies in Statistics 6 I
ÖREBRO 20132013
DA
O L
I
Co
m
m
on F
ea
tu
re
s i
n V
ec
to
r N
on
lin
ea
r T
im
e S
eri
es M
od
els
dao li received her Bachelor degree in Computer Science
& Technology from Communication University of China, Beijing 2004. She started her graduate studies in Statistics at Dalarna University, Sweden 2007, after two years of wor-king as a software engineer in a China-based international company, which is at the forefront of the mobile solution designs. She received her Licentiate of Philosophy degree in Statistics from Örebro University, Sweden 2011. In the latest years she has been devoted to studying econometrics, developing contributions related to hypothesis testing and forecasting, teaching courses and supervising essays in Statistics.
Her main research interest is nonlinear time series econometrics. The work presented in this thesis focuses on testing, modeling and forecasting nonlinear common features. Both stationary and nonstationary time series are concer-ned. A definition of common features is proposed in an appropriate way to each class. Based on the definition, a vector nonlinear time series model with common features is set up for testing for common features, and a testing procedure is provided. Tests are derived and their finite-sample properties are studied by Monte Carlo simulations. After specified using the testing contri-butions above, the model is exemplified in detail for forecasting with vector nonlinear models, and illustrated by an application of macroeconomic data.
issn 1651-8608 isbn 978-91-7668-952-3