A note on estimation by least squares for harmonic component models
Let observations (X_1, H ,X_n) be generated by a harmonic model such that X_t=A_0 cos omega_0t + B_0 sin omega_0t + epsilon _t, where A_0,B_0, omega_0 are constants and ( epsilon _t) is a stationary process with zero mean and finite variance. The estimation of A_0,B_0, omega_0 by the method of least squares is considered. It is shown that, without any restriction on omega in the minimization procedure, the estimate is an n-consistent estimate of omega_0, and hence ( ) has the usual asymptotic distribution. Copyright 2003 Blackwell Publishing Ltd.
Year of publication: |
2003
|
---|---|
Authors: | Walker, A. M. |
Published in: |
Journal of Time Series Analysis. - Wiley Blackwell, ISSN 0143-9782. - Vol. 24.2003, 5, p. 613-629
|
Publisher: |
Wiley Blackwell |
Saved in:
Saved in favorites
Similar items by person
-
A note on the relation between intelligence and parental income
Walker, A. M., (1953)
-
A reexamination of some basic asymptotic theory for linear processes in time series analysis
Walker, A. M., (1983)
-
A note on the asymptotic efficiency of an asymptotically normal estimator sequence
Walker, A. M., (1963)
- More ...