Intelligent Systems
Take Home Quiz 1 - Due Feb 11, 1998
1. Use a modified version of the Matlab Neural Network Toolbox program DEMOP1 to correctly classify the Bongard data. The data as measured in class is given in the previous course notes. You must modify the target vector, T, to make a two parallel perceptron classifier that can accomplish the XOR classification. A sample program segment is shown below.
% A row vector T defines the vector's target categories.
% A two-row target vector defines the two
% parallel perceptron neural net that performs the XOR classification
T = [1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 0 0 1 1 0 0 0 0 0 ;
0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 1 0 0 0 0 0 ];
2. Determine whether either of the following two points:
P1 = [3/4 7/4 5/4 7/4]T
and
P2 = [5/4 9/4 7/4 7/4] T
is on the line through
Q1 = [1 2 1 2] T and Q2 = [2 3 4 1] T
3. Determine whether the following two lines are parallel, where
L1 = { x: x = [2 3 1] T + a [1 4 -1] T }
and L2 = {x: x = [5 7 2] T + a [ -3 -12 3] T}
4. Let L be the line through
P1 = [1 2 1 3] T and P2 = [ 2 3 4 1] T
Find the equation of a line L2 through P1 and perpendicular to L1.
H1 = { x: [4 2 7 8 ] x = 4} and H2 = { x : [1 1/2 7/4 2] x = 19}
determine the point y on H closest to z.
Sample number |
Measurement |
Class |
1 |
0.608 |
2 |
2 |
-1.590 |
1 |
3 |
0.235 |
2 |
4 |
3.949 |
2 |
5 |
-2.249 |
1 |
6 |
2.704 |
2 |
7 |
-2.473 |
1 |
8 |
0.672 |
2 |
9 |
0.262 |
2 |
10 |
1.072 |
2 |
Week |
Stock Price |
1 |
37.25 |
2 |
36.875 |
3 |
36.625 |
4 |
37.625 |
5 |
39.75 |
6 |
38.75 |
7 |
39.125 |
8 |
38.875 |
9 |
39.375 |
10 |
42.25 |
11 |
43.125 |
12 |
41.375 |
13 |
41.125 |
14 |
40.25 |
15 |
39.0 |
16 |
38.375 |
17 |
38.25 |
18 |
39.25 |
PLEASE TURN IN YOUR SOLUTIONS IN SUFFICIENT DETAIL TO PERMIT A VERIFICATION OF YOUR WORK. A TYPED SOLUTION IS PREFERRED. A CAREFULLY PRINTED HANDWRITTEN SOLUTION IS ACCEPTABLE. PROGRAM LISTINGS FOR PROBLEMS 1,7 AND 8 ARE NECESSARY AS WELL