Research On Several Types Of Almon Estimators In Finite Distributed Lag Model

The finite distribution lag model is a common time series model,and the study of parameter estimator in this model is of great significance.A common estimator method in this model is the Almon-least squares estimator,which is proposed by combining the Almon polynomial lag method with the least squares estimator.In order to solve the problem of explaining the complex collinearity of variables,relevant scholars have also proposed some alternative metrics for Almon-least squares estimator,such as Almon-ridge estimator.Based on the finite distribution lag model,three new Almon-biased estimators are proposed by combining the Almon polynomial lag method with three kinds of biased estimators.The predictive mean square error matrix or the predictive mean square error is used as the evaluation criterion.The predicted performance of the new estimator was studied.Firstly,the application of the finite distribution lag model,the research background,the research status at home and abroad,the research innovation and the deficiencies are introduced.Secondly,the model,Almon polynomial lag method,evaluation criteria and related lemma are introduced.The Almon-two-parameter estimator is obtained by combining two-parameter estimator and Almon polynomial method.Under the criterion of predictive mean square error matrix and predictive mean square error,the predictive superiority of Almon-two-parameter estimator with respect to Almon-least squares estimator,Almon-ridge estimator,and Almon-Liu estimator was studied.The Almon-principal component two-parameter estimator is proposed by combining the Almon polynomial lag method and the principal component two-parameter estimator.The predictive superiority of Almon-principal two-parameter estimator is compared with the Almon-least squares estimator,Almon-r-k class estimator,Almon-r-d class estimator in the sense of the predictive mean square error matrix or the predictive mean square error.Based on the Almonpolynomial lag method and the principal component Liu type estimator,the Almon-principal component Liu type estimator is obtained.The Almon-principal component Liu type estimator is compared with the Almon-least squares estimator,Almon-principal component estimator in the sense of predictive mean square error matrix or the predictive mean square error.Finally,we summarize the content of this paper and put forward some thoughts and suggestions on the work to be carried out in the next step.