Lecture on AIC: compare non-nested
models
What was covered?
Using AIC to compare models in which some, but not all, of the
models use a transformed response variable
Models for excess zeros
AIC for comparing models when some, but not all, use a transforme
A Handbook of Statistical Analyses
Using R
Brian S. Everitt and Torsten Hothorn
CHAPTER 6
Logistic Regression and Generalised
Linear Models: Blood Screening,
Womens Role in Society,
and Colonic Polyps
6.1 Introduction
6.2 Logistic Regression and Generalis
University of North Carolina at Charlotte
STAT7133-8133 Midterm
Spring 2012
1. Briefly answer the following questions:
(a) Suppose that the p-dimensional random vector X has mean and covariance
)
matrix , which is positive definite. Then (
(
) is distribu
Use the following worked example as take-home part. You need to work it out by hand, but not by
software. To get a full credit, you need to show your work in detail.
Lecture Note: Other link functions and Goodness-of-fit test
What was covered?
Other link functions used in generalized linear models
The deviance as a goodness of fit statistic
Analysis of deviance
Terminology Defined
probit link
complementary log log lin
University of North Carolina at Charlotte
STAT7127-8127
Midterm
Fall 2009
Nov. 2, 2009
PLEASE PRINT THE FOLLOWING INFORMATION:
Name: _
Student ID#: _
Instructor: Jiancheng Jiang
Time: 11:00 am 12:15 pm
This test consists of two parts. Part I is take-home,
Lecture 11 - STAT 3150
Order Determination
To introduce general principles for determining the
order (p,q) in ARMA modeling: AIC, BIC and FPE.
(i)
Akaike Information Criterion (AIC)
AIC is used to select the optimum parametric
model based on observed dat