STAT 760, September 23, 2011
Homework 1, Due 9/28/11
In the system under study we are able to observe pdimensional random vectors Y1 , Y2 , . . . , YT . We
model them as follows:
Yt = F t + Vt
(1)
for t = 1, 2, . . . , T , where F is a known p q matrix, p
STAT 760, November 15, 2011
Homework 3, Due November 23
1. Summarize your proposed nal project (about 1 page). Email this to Prof Newton separately
from the other homework solutions.
2. Linear Discriminant Analysis (LDA-1): Apply LDA-1 to build a classier
STAT 760, October 3, 2011
Lecture on principal components
Setup: See text about perceived utility of dimension reduction schemes. They are widely used,
though sometimes dicult to interpret, and not so easily motivated from any rst-principles on the
dynami
STAT 760, October 31, 2011
Name:
1. Consider a nite undirected graph G and a set of random variables X = cfw_Xv indexed by
vertices of G. What does it mean for X to satisfy the pairwise Markov property with respect
to G? Similarly, what does it mean to s
STAT 760, November 28, 2011
Name:
1. Consider an n p data matrix X holding measurements on n experimental units and p
variables. Carefully describe the steps involved in a hierarchical cluster analysis that tries to
cluster the experimental units.
2. Cons