Strategies for Variable Selection
Chp 12 of The Statistical Sleuth
Last Updated: 24/08/2016
Model Selection Among all Subsets
These methods fit all possible candidate models. A value of the criterion
Strategies for Variable Selection
Chapter 12 of The Statistical Sleuth and STAT 2008 notes.
Last Updated: 21/08/2016
There are two prime reasons for variable selection:
1. Simple models are preferable
Tests and Confidence Intervals for Linear Combinations of Coefficients
Take from Chapter 10 of The Statistical Sleuth
Last Updated: 15/08/2016
1
Inference about the mean at some combination of Xs
For
MUTLIPLE LINEAR REGRESSION (MLR)
Taken from CHAPTER 9 of The Statistical Sleuth
Last Updated: 04/08/2016
MLR models the mean of the response variable as a function of several
explanatory variables. Ex
SIMPLE LINEAR REGRESSION
Taken from CHAPTER 7 of The Statistical Sleuth
Last Updated: 25/07/2016
At the end of last lecture, we looked at the sampling distributions of b0 and b1.
To use these equation
Note: The lecture notes for this course are primarily based on the class text The Statistical Sleuth by Fred L. Ramsey and Daniel W. Schafer.
These notes summarise/highlight the material that is prese
Inferential Tools for Multiple Regression
Taken from Chapter 10 of the Statistical Sleuth
Last Updated: 06/08/2016
For SLR we saw that = (Y Y ) . For MLR we have a similar result:
2
i
i
n2
= n(YpY )
Model Checking and Refinement
Taken from Chapter 11 of The Statistical Sleuth and STAT 2008 lecture notes.
Last Updated: 21/08/2016
Partial Residual Plots
Partial residual plots help us decide how, an
SLR ASSUMPTIONS
Taken from Chapter 8 of The Statistical Sleuth
Last Updated: 01/08/2016
Most of the inferential tools of the previous two lectures relied heavily on the
SLR assumptions: linearity, con
R-Squared
Taken from Chapter 8 of The Statistical Sleuth and ANU STAT2008 lecture notes
Last Updated: 07/08/2016
Partitioning Variability
In any dataset, there will be variation in the values of the r