linearRegression5_4perPage

# linearRegression5_4perPage - 3 Just H2S We really should...

This preview shows pages 1–2. Sign up to view the full content.

Multiple Linear Regression (Last) Data Mining Prof. Dawn Woodard School of ORIE Cornell University 1 Outline 1 Announcements 2 Multiple Linear Regression 2 Announcements Questions? 4 C p Statistic Obtain the cheese data from Blackboard Read the data into R: cheese <- read.table( "C:/temp/cheese.txt", skip = 8, col.names = c("ID", "taste", "acetic", "h2s", "lactic")) p = ? N = ? How many models to consider? What are they? Fit the full linear model; summarize the results summary( cheeseLM ) Look at the residuals ( Y i ˆ Y i ) : cheeseLM \$ residuals 6

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
C p Statistic Calculate the RSS for the full model. What is the code? Fit the following models, and for each ±nd the RSS: 1 H2S and Lactic 2 Acetic and H2S
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 3 Just H2S We really should consider all the models, but to save time I have just picked the promising ones. 7 C p Statistic Write down the RSS for each model: RSS Full model H2S & Lactic Acetic & H2S Just H2S Calculate the C p statistic for each model: C p Full model H2S & Lactic Acetic & H2S Just H2S What model do we choose? 8 C p Statistic What is the value of C p for the full model? 9 Model Choice Criteria There are several criteria for model choice that do not use a test data set E.g., C p , Aikake’s Information Criterion (AIC), Deviance Information Criterion (DIC) 10...
View Full Document

## This note was uploaded on 12/23/2009 for the course ORIE 4740 at Cornell.

### Page1 / 2

linearRegression5_4perPage - 3 Just H2S We really should...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online