From "Practical Regression and ANOVA
Using R" by Julian Faraway, 2002
Chapter 10
Variable Selection
Variable selection is intended to select the best subset of predictors. But why bother?
1. We want to explain the data in the simplest way redundant predic
Data Analysis Math 7360
Spring 2014
The manipulation of statistical formulas is no substitute for knowing what one is doing.
~Hubert M. Blalock, Jr.
Instructor: Michelle Lacey
Office: Gibson 417C
Phone: (504) 862-3439
E-mail: mlacey1@tulane.edu
Office Hou
Robust Regression
Appendix to An R and S-PLUS Companion to Applied Regression
John Fox
January 2002
1
M -Estimation
Linear least-squares estimates can behave badly when the error distribution is not normal, particularly when
the errors are heavy-tailed. O
Math 7360, Spring 2014
Data Analysis
Page 1 of 2
Handout #7
Robust Regression
As we know, standard linear regression finds coefficients to minimize the sum of squared
2
differences between the fitted values and the response, y i y i . While this is genera
Math 7360, Spring 2014
Data Analysis
Page 1 of 2
Handout #8
Generalized Linear Models
Models for Binary Response Data
Logistic regression methods are typically employed for binary response data, implemented in R
using the glm function. Our dataset may spe
Math 7360, Spring 2014
Data Analysis
Page 1 of 2
Handout #10
Nonlinear Least Squares
In some cases, we wish to estimate parameters for models that are nonlinear in nature, such as
models for population growth or weight loss. If we have some function that
Math 7360, Spring 2014
Data Analysis
Page 1 of 3
Handout #3
Lesson 2 (cont): Looking at data, univariate analysis
Density Estimation
Although a relative frequency histogram provides a sort of estimate of the density function for a
dataset, it only places
Math 7360, Spring 2014
Data Analysis
Page 1 of 3
Handout #4
Lesson 3: Linear regression
A regression model expresses one variable as a linear combination of one or more other
variables. We call the fitted variable the response, and the other variables are
Math 7360, Spring 2014
Data Analysis
Page 1 of 4
Handout #2
Lesson 1 (cont.): Working with numbers and generating data.
Distributions
For all of the standard statistical distributions, R includes functions to compute probabilities,
densities, and quantile
Math 7360, Spring 2014
Data Analysis
Page 1 of 7
Handout #1
A Crash Course in R
Lesson 0: Motivation.
This package may not seem too appealing at first, because its not particularly user-friendly.
You might think to yourself, Why cant we just work with Exc
Math 7360 Homework Assignment #2
The R dataset Cars93, stored in the MASS library, provides records for 93 cars selected at
random from among 1993 passenger car models that were listed in both Consumer Reports and
the PACE Buying Guide. The variables incl
Math 7360, Spring 2014
Data Analysis
Page 1 of 2
Handout #6
Lesson 5: Model Selection
Lets load the dataset mtcars. This dataset was extracted from the 1974 Motor Trend US
magazine, and contains 11 variables on aspects of automobile design and performance
Math 7360, Spring 2014
Data Analysis
Page 1 of 7
Handout #5
Lesson 4: Data types and classes, modeling with non-numeric data
An object in R is assigned a class attribute, which describes the object and determines which
functions are applicable to it. For