NOTES for Thursday 2012.MAR.29
Now ready for the extension to multiple regression. It looks like we just
expand the situation to cover multiple predictors (independent variables). The
data will usually come in a spreadsheet layout. The variab
NOTES for Wednesday 2012.MAR.21
The regression problem does not treat the two variables in the same way.
One of the variables is called the independent variable, and its given the
horizontal axis. It plays the role of x (but of course it may
NOTES for Monday 2012.MAR.19
This first example illustrates the normal random walk for equity prices. This was not
done in class. Most analyses of equity prices use the lognormal random walk, which
comes a little later.
Suppose that we have a
NOTES for Thursday 2012.MAR.08
Some fun, easy math.
Lets say you get a 10% raise, then you get a 10% pay cut. Are you better off, worse off,
or the same?
Suppose that your base salary is B. The actual value of B does not matter.
The 10% raise
NOTES for Wednesday 2012 MAR 07
And now. back to stories about the sample average X .
Recall that in the sampling story we had E( X ) = and also SD( X ) =
Major use is law of averages. This says simply X converges to , the population mean
NOTES for Monday, 2012.MAR.05
Suppose that X1 , X2 , , Xn is a sample from a population with mean and standard
. Of course, the sample standard deviation s
deviation . We know that SD ( X ) =
is just a data-based estimate of . The data-ba
Notes for Thursday 2012.MAR.01
Heres a quick review about what weve done with random variables.
For discrete random variables, we studied the binomial, the Poisson, and the
hypergeometric. We should know the probability functions, along with
NOTES for Monday 2012.JAN.30
By the way, if the list x1, x2, , xn constitutes an entire population (rather than a
sample) we define the standard deviation as =
xi , where we use
xi as a replacement for x . For now, do not be over
NOTES for Thursday 2012.JAN.26
Weve been dabbling in various numeric summaries for quantitative data. Crudely we
separate these (so far) in terms of measures of location (position) and measures of spread
The common statistical
NOTES for Wednesday 2012.JAN.25
Lets consider a number of devices for displaying quantitative, or measured, data.
These devices can also be used for ordered qualitative data.
We can always make histograms for quantitative data. Lets illustrat
NOTES for 2012.JAN.23
Office hours for Gary Simon, Spring 2012
These are tentative, pending committee times.
3:30 p.m. 5:00 p.m.
noon 1:30 p.m.
The course syllabus has been distributed in class. Please note the important dates.
NOTES for 2012.FEB.29
Lets note some interesting things you can do in Minitab with regard to continuous
The command Graph Probability Distribution Plot will allow you to see various
probability densities. Here, for example, is
NOTES for Monday, 2012.FEB.27
We need a way to describe continuous random variables. There is no probability at any
single value, so we cant simply use descriptions of the form P[ X = k ].
You can think about this in terms of questions like W
NOTES for Wednesday 2012.FEB.15
Suppose that a hospital has a cardiac care unit which handles heart attack victims
on the first day of their problems. The geographic area served by the hospital
creates new patients as observations on a Poisso
NOTES for Monday 2012.FEB.13
If X is a binomial random variable with parameters n and p (meaning that X is the
number of successes in n independent trials where each trial has success
probability p), then the mean of the distribution of X is
NOTES for Wednesday 2012.FEB.08
Here is an interesting misuse of independence. The website
http:/howmanyofme.com claims to give the number of people in the United
States with any name. They do this for first names, last names, and then for fu
Notes for Monday 2012.FEB.06
Reprise on the taxi problem. Let
be the event that a green taxi caused the accident
be the event that a blue taxi caused the accident
be the event that Ralph said green
be the event that Ralph said blue
Notes for Wednesday, 2012.FEB.01.
For many problems, we need to do a little counting. We try to construct a sample
space S for which the elements are equally likely. Then for any event E, we will
have P ( E ) =
elements in E
. It can tricky c
NOTES for Thursday 2012.APR.05
Advantages of best subsets regression and stepwise regression:
The procedures are automated, so that the user does not have to think about
correlations, VIF numbers, residual sums of squares.
The procedures actu
NOTES for Wednesday 2012.APR.04
One of the clues on the library data comes through the VIF values. These VIFs tell you to
what extent a predictor is linearly dependent on other predictors. We like these to be
close to 1, and we certainly get
NOTES for Monday 2012.APR.02
Multiple regression creates comes some difficulties with the interpretation of
coefficients. Suppose that you do the regression of Y on (G, H, K). The
estimated coefficient bG represents the association between va