The Theoretical Study Of Integrated C-K Conditional Ridge Estimation

Introduce appropriate biased estimation to improve the least squares estimation of the defects of complex collinearity problem, in the absence of constraints of linear regression model in the study of the development of the already quite mature. But in a large number of practical research problems, the parameter is often accompanied by some constraints to increase difficulty to the research process, but also very meaningful. Under which the equality constraint model under the constraints of the squares estimate also suffer complex collinearity problem, also need to use biased estimation to improve insufficient. The current study is not confined to between the estimation and least squares estimate optimal benign comparison, also attaches great importance to between all kinds of biased estimation optimal benign comparison, it is also very has the theoretical and practical significance.This paper mainly studies in unconstrained biased estimation based on further complement, and is applied to the theoretical thoughts constraint biased estimate of the study, mainly do the following several aspects work:(1) for integrated c-K ridge estimation on the issue of ridge parameters selection, by constructing the estimated mean square error of the unbiased estimation of the unbiased estimation of minimizing the mean square error solution as a kind of method.(2) Put forward integrated c-K conditional ridge estimation for the regression model with homogeneous equation constraints, and given the nature of the multiple show its rationality. And focus on the estimation and least squares estimate, and other constraints biased estimate of the optimal benign comparisons, it is concluded that the estimate of necessary and sufficient condition is more superior and partial parameters scope.(3) In estimate the choice of partial parameters, using the iterative algorithm to discuss estimates of iterative solution.