Estimating Distributed Lag Coefficients when there are Errors in the Observed Time Series.
Abstract
Estimating the distributed lag coefficients (h(mTAU)) from a sample of the two processes when (x(NTAU)) and (y(nTAU)) are measured with error is a statistical problem that is frequently encountered in physical science, engineering, and social science applications. In the engineering and science literature the distributed lags are called the impulse response weights of a causal linear filter. A least squares fit of the model gives biased estimates of the coefficients for this time series version of the errors-in-variables problem. This paper presents approximately unbiased estimators of a scalar multiple of the coefficients. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1982
- Accession Number
- ADA123460
Entities
People
- Melvin J. Hinich
Organizations
- Virginia Tech