Prediction of Future Observations in Polynomial Growth Curve Models. Part 1.
Abstract
The problem considered is that of simultaneous prediction of future measurements on a given number of individuals using their past measurements. Assuming a polynomial growth curve model, a number of methods are proposed and their relative efficiencies in terms of the compound mean square prediction error (CMPSE) are compared. There is a similarity between the problem of simultaneous estimation of parameters as considered by Stein and that of simultaneous prediction of future observations. It is found that the empirical Bayes predictor (EBP) based on the empirical Bayes estimator (EBE) of the unknown vector parameters in several linear models proposed by the author (Rao, 1975) has the best possible efficiency compared to the others studied. The problem of determining the appropriate degree of the polynomial growth curve is also studied from the point of view of minimising the CMSPE. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1983
- Accession Number
- ADA129359
Entities
People
- Calyampudi Radhakrishna Rao
Organizations
- University of Pittsburgh