HP 15c - Manual (Page 114)

114 Section 4: Using Matrix Operations
114
You will need the original dependent variable data for each regression. If there is not enough
room to store the original data in matrix E, you can compute it from the output of the
regression fit. A subroutine has been included to do this.
This program has the following characteristics:

If the entire program is keyed into program memory, the sizes of n and p are required
to satisfy n ≥ p and (n + p)(p + 1) ≤ 56. That is,
if p is 1 2 3 4
then n
max
is 27 16 11 7
This assumes that only data storage registers R
0
and R
1
are allocated. If subroutine
"B" is omitted, then n ≥ p and (n + p)(p + 1) ≤ 58. That is,
if p is 1 2 3 4
then n
max
is 28 17 11 7
 Even though subroutine “B” uses the residual function with its extended precision,
the computed dependent variable data may not exactly agree with the original data.
The agreement will usually be close enough for statistical estimation and tests. If
more accuracy is desired, the original data can be reentered into matrix B.
Keystrokes Display
|¥
Program mode.
´CLEARM 000-
´bA 001-42,21,11 Program to fit model.
l>B 002-45,16,12
´>8 003-42,16, 8
|x 004- 43 11 Calculates Tot SS.
l>A 005-45,16,11
v 006- 36
´ 007-42,26,13
´>5 008-42,16, 5 Calculates C = A
T
A.
|K 009- 43 36
l>B 010-45,16,12
´<Á 011-42,26,14
´>5 012-42,16, 5 Calculates D = A
T
B.
® 013- 34
÷ 014- 10 Calculates parameters in D.
l>A 015-45,16,11
® 016- 34
´ 017-42,26,12