Font Size: a A A

New Estimates And Combined Tests Of Coefficients In Regression Models

Posted on:2021-04-30Degree:MasterType:Thesis
Country:ChinaCandidate:L S YinFull Text:PDF
GTID:2430330623471406Subject:Probability theory and mathematical statistics
Abstract/Summary:PDF Full Text Request
With the rapid development of electronic technology,the regression model are often generated by repeated measurement test,multistage sampling survey,and with time and the individual effect of economic research.The Panel model widely used in econometrics,regional economic research,market analysis,etc.The regression model of the related content has important practical signifi-cance,this article focus on estimation and combination of regression coefficient in the model test are discussed.The main research content is the following two parts:The second chapter to contain multiple linear regression model in the total coefficient esti-mation and ridge parameter determination were studied.The linear model is a commonly used statistical model.When there are multiple collinearity between the independent variables,the com-monly used methods of parameter estimation is ridge estimate and principal components estimate,based on the ideas of the ridge estimate and principal components estimate,a new estimation method is presented.In the sense of mean square error,and ridge estimate,principal components estimates,compared in this article to estimate has a smaller mean square error.At the same time,this paper discusses the determination of ridge parameters is given in theory.The third chapter,the linear hypothesis test of regression coefficient in the model with three random effects and its simulation effect are discussed and studied.Under the balanced Panel da-ta model with three random effects,by constructing exponential statistics to design the accurate test of Panel model,the accurate test method of linear hypothesis test is given.On this basis,a numerical example is given to prove the correctness of the conclusion.
Keywords/Search Tags:Linear regression model, Ridge estimation, Principal component estimation, Mean square error, Panel data model, Precise inspection
PDF Full Text Request
Related items