Analysis of seasonality in restricted least squares method with dummy variables and an application

A priori information on some of the relations between the coefficients ofregression may be available. This a priori information is received from thesample, and in the precense of a priori information in order to reach efficientestimation of the parameters, Restricted Least Squares (RLS) method isemployed. In this thesis RLS is investigated under the precense of LinearRestrictions (constraints) and the importance of the tests about the validity of therestrictions are introduced. On the other hand Mean Square Error Kriteria ishandled becouse of its importance in the method of RLS. While the estimationswere carried out under the assumption that X matrix in the models which wasestablished in last section (Application) of the thesis is full rank, new approacheson the estimation When X matrix is not of full rank, is also mentioned. These newapproaches especially used dummy variables, and seasonal variables are definedas dummies in regression analysis.The quarterly data of the Turkish Labour Market in the period between1999-2005 is held and RLS method is applied. The precense of seasonality in themodels which are constructed by dummy variables with different trends aredetermined with RLS methodKeywords: A priori Information, Linear Restrictions, Restricted Least SquaresMethod,Dummy Variables, Seasonality.