A Variable Selection Procedure for Two Stage Least Squares Method

スポンサーリンク

概要

• 論文の詳細を見る

• The variable classifications based on the characteristics of the two stage least squares(2SLS) and scientific knowledge of the research field lead to the derivation of meaningful and just- or over-identifiable subsets from a set of all possible included predetermined, explanatory endogenous and excluded predetermined variable candidates specified for an explained endogenous variable. The j-th best subset is defined as the one which is meaningful for the research, is just- or over-identified, passes the criteria set for the signs and/or magnitudes (of values or absolute values calculated by linear functions) of estimated coefficients, the Basmann over-identifiability restriction test, the Durbin-Watson serial correlation test, the absolute relative error and the turning point tests for estimated observations, and has the j-th highest adjusted coefficient of determination. The first to the J-th best subset, for example, J=10, can be chosen from all possible subsets in one computer-run by the package OEPP.

• 1986-01-30

著者

• Onishi Haruo

  Institude Of Policy And Planning Sciences University Of Tsukuba

関連論文

• LIMITED INFORMATION MAXIMUM LIKELIHOOD USING PRINCIPAL COMPONENTS AND ITS APPLICATION
• A KNOWLEDGE-BASED VARIABLE SELECTION METHOD FOR BOX-COX TRANSFORMATION
• A Variable Selection Procedure for Two Stage Least Squares Method
• An Efficient Method for Building a Prefectural Government Official Reduction Model by the Researcher System OEPP
• Box-Cox Transformation and Informatic Variable Selection

スポンサーリンク