Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Problem computing eigenvalues, STATA vs. MATA
From
Steven Samuels <[email protected]>
To
[email protected]
Subject
Re: st: Problem computing eigenvalues, STATA vs. MATA
Date
Sun, 9 Jan 2011 20:44:44 -0500
Reread Austin's response and then look up the definition of
eigenvalues and eigenvectors. The two "different" solutions are, in
fact, equivalent.
Steve
[email protected]
On Jan 9, 2011, at 5:30 PM, Zachary Neal wrote:
This doesn't seem to help. Running the code suggested below by Austin
Nichols, the values in the first eigenvector using matrix symeigen
are:
.5773502692
.5
.5
.2886751346
.2886751346
However, the values in the first eigenvector using
symeigensystem(st_matrix("A"),vectors2,values2) are:
.5773502692
-.5
-.5
.2886751346
.2886751346
I'm not clear why the 2nd and 3rd values here are not positive, as
they are in the results produced by matrix symeigen.
Best
Zachary
On Sun, Jan 9, 2011 at 10:42 AM, Austin Nichols
<[email protected]> wrote:
Zachary Neal <[email protected]>:
You are confusing vectors and values there, and not sorting by values
after computing the eigensystem, but note that the only real
difference is that a couple of vectors are negated (any scalar
multiple of an eigenvector is also an eigenvector and the negation has
the same length so they are essentially equivalent) and Mata is using
quad precision to do the computations.
matrix define A = 0,1,1,0,0 \ 1,0,0,1,0 \ 1,0,0,0,1 \ 0,1,0,0,0 \
0,0,1,0,0
matrix symeigen vectors1 values1 = A
matrix list values1
matrix list vectors1
mata
values2 = 0
vectors2 = 0
symeigensystem(st_matrix("A"),vectors2,values2)
v2=sort((values2\vectors2)',1)'
v2[1,.]
v2[2..6,.]
end
On Sun, Jan 9, 2011 at 8:08 AM, Zachary Neal <[email protected]> wrote:
I am trying to obtain the eigenvalues of a symmetric matrix.
However,
I get different results depending on whether I use STATA or MATA.
For
example:
matrix define A = 0,1,1,0,0 \ 1,0,0,1,0 \ 1,0,0,0,1 \ 0,1,0,0,0 \
0,0,1,0,0
matrix symeigen eigenvalues1 eigenvectors1 = A
matrix list eigenvalues1
mata: A = 0,1,1,0,0 \ 1,0,0,1,0 \ 1,0,0,0,1 \ 0,1,0,0,0 \ 0,0,1,0,0
mata: eigenvalues2 = 0
mata: eigenvectors2 = 0
mata: symeigensystem(A,eigenvalues2,eigenvectors2)
mata: eigenvalues2
In this case, eigenvalues1 does not equal eigenvalues2. I believe
the
results yielded by STATA (i.e. eigenvalues1) are what I'm looking
for.
Why do these two sets of commands yield different results? What
commands are necessary in MATA to yield the same eigenvalues that are
given by STATA?
Thank you
Zachary Neal
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
--
Zachary Neal, PhD
Department of Sociology
Michigan State University
[email protected]
http://www.msu.edu/~zpneal
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/