This preview shows pages 1–7. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Lecture 4 : Ordinary Least Square(OLS) Econ 444, Fall 2010 Jan 21, 2010 The Estimated Regression Consider the regression equation, Y i = + 1 X i + i i = 1 , 2 ...., N (1) The problem is in real life we only observe a sample and hence the population parameters and 1 are unknown. The regression coefficients and 1 need to be estimated from a data set (or a sample). We denote the estimators by and 1 , respectively.Then, Y i = + 1 X i (2) is called the estimated (or sample) regression . Y i is called the estimated or fitted value and and 1 are called estimated regression coefficients . Estimating a regression using OLS In this class we discuss how to estimate the unknown parameters and 1 in the regression equation, Y i = + 1 X i + i (3) This is equivalent to fitting a straight line to the data. Visually inspecting the data and fitting a line is not a good idea as it is too arbitrary and imprecise. We choose and 1 in such a way that the residual e i = Y i Y i = Y i ( + 1 X i ) (4) is as "small" as possible under a certain criterion. There are many criteria that one could use but for this course we will use the the Ordinary Least Squares method. The estimators of and 1 based on this method are called OLS estimators are denoted by and 1 respectively. Algebraically, the least squares principle tells us to choose and 1 , so that we minimize residual sum of squares (RSS) N X i = 1 e 2 i = N X i = 1 ( Y i ( + 1 X i ) 2 (5) by choosing and 1 appropriately. How Does OLS Work?...
View Full
Document
 Winter '07
 OGAKI
 Econometrics

Click to edit the document details