MATH 226 - 10.3.2007

# MATH 226 - 10.3.2007 - various variables and their levels...

This preview shows page 1. Sign up to view the full content.

MATH 226 – October 3, 2007 Regression – object to build a model, such that observations y can be predicted from x (x 1 , x n ) Comparison of effects o (1) ŷ= α 1 + β 1 xif β 1 > β 2 x has bigger influence on y (2) ŷ= α 2 + β 2 x in model 1 than in 2 o All variables are qualitative Measures of Accuracy o Should be known o R 2 = (Σ(y i – y-bar) 2 - Σ(y i – ŷ i ) 2 )/ Σ(y i – y-bar) 2 o Coefficient of Determination 0 ≤ R 2 ≤ 1 o r correlation coefficient only tells about linear relation between two variables o Residual Plots There should be no pattern Vs. ŷ (fitted value) vs. y vs. x 1 :x n Fitted effects for factorial experiments: some or all of the x’s are qualitative The objective is not to predict, rather to explain the influences of
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: various variables and their levels to the response o The variables x 1 :x n are called factors and their values are levels Fitted main effects are defined as the overall difference of a level of the factor as compared to the average o A, B, C are the factors o So the fitted main effect at level i of factor A is a i = y i• bar - y •• bar Fitted interaction effects One goal is to fit the simplest model o μ + α i + β j + αβ ij μ + α i μ + β j Measure of Fit o R 2 o Residuals (y – ŷ) o Absolute value size comparison of effects...
View Full Document

## This note was uploaded on 04/07/2008 for the course MATH 226 taught by Professor Daepp during the Fall '07 term at Bucknell.

Ask a homework question - tutors are online