| | |
Summary: Assignment #2
By now, you should be comfortable with data les in general, and with
plotting.
You are, here, asked to generate some arti cial data according to some
parameters, then see whether you can extract those parameters back from
the data. Also, you will construct some simple, two-category classi ers.
Your programming systems may have a generator for normal distribu-
tions, in which case use that. If it doesn't, you can approximate bell curves
by using a uniform generator (call it \R()") that produces numbers in the
range [ 0:5; +0:5]. Then, approximate a normal (bell-shaped) distribution
by summing ve terms:
R() +R() +R() +R() +R()
The last questions ask you to apply the machinery developed for arti cial
data to real data.
1. Generate and plot two-dimensional normal scatter diagrams of various
sizes. Verify that the center of mass of the clouds is nearly (0; 0).
2. Call such a cloud of data C. Generate and plot other clouds C 0 as
follows. Select a 2 2 matrix, M , and a vector, ~
w.
C 0 = fM~v + ~
|
