scientific article; zbMATH DE number 3963013
The classical approach for estimating unknown parameters in a linear regression model, namely the least squares estimator, has got many competitors in recent years. Its disadvantages have been discussed in quite a few papers. In this booklet the author investigates properties of generalized ridge estimators. One should note that this booklet is a typical Ph.D. dissertation, i.e. you will find many quotations, various results the proof of which has to be found in the literature, there is no subject index (but a very good list of references) and the type of presentation is often in terms of stating theses which have to be defended or attacked. The mathematics used is clear and on the usual level of an exposition in applied statistics. The main new contribution is in the reviewer's opinion a thorough discussion of the ridge trace and a noteworthy proposal to measure its (in-)stability. This measure is the basis for a simulation study which fills sixty pages and has the result (in the interpretation of the reviewer) that none of the many open problems in this area can be solved by the computers but that they may provide hints for further theoretical research.
- Estimation of the signal-to-noise in the linear regression model
- Variations on ridge traces in regression
- Mean squared error matrix comparisons between biased estimators — An overview of recent results
- The general linear test in the ridge regression
- Comparison of regression estimators using Pitman measures of nearness.
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3730869)