Numerical Analysis

Selecting certain columns of Matrix

Hi,

Can you think of some "selection matrix" to select only certain colum
vectors of a matrix?
For example,
X=[x1 x2 x3 x4] where xi is a column vector.
I would like to have a matrix J such that
XJ or JX will be [x1 x3] by selecting only 1st and 3rd columns of the
matrix.

Thanks in advance.

.
posted by admin in Uncategorized and have Comments (6)

6 Responses to “Selecting certain columns of Matrix”

  1. admin says:

    On Jan 20, 2:41 pm, usuiisu <ikuy…@gmail.com> wrote:

    > Hi,

    > Can you think of some "selection matrix" to
    > select only certain colum vectors of a matrix?
    > For example,
    > X=[x1 x2 x3 x4] where xi is a column vector.
    > I would like to have a matrix J such that
    > XJ or JX will be [x1 x3] by selecting only
    > 1st and 3rd columns of the matrix.

    > Thanks in advance.

    To get columns 1 and 3, postmultiply X by
    J = {{1,0},
         {0,0},
         {0,1},
         {0,0}}.

  2. admin says:

    On Thu, 20 Jan 2011 14:41:19 -0800, usuiisu wrote:
    > Hi,

    > Can you think of some "selection matrix" to select only certain colum
    > vectors of a matrix?
    > For example,
    > X=[x1 x2 x3 x4] where xi is a column vector. I would like to have a
    > matrix J such that XJ or JX will be [x1 x3] by selecting only 1st and
    > 3rd columns of the matrix.

    That depends on the programming language you’re using.

    In Scilab (Matlab) X[(1,3,7)] is a matrix with the 1st, 3rd, 7th
    column of the original matrix X.

    In C++ you can use the "Eigen" matrix package (IMHO, the best (free) one)
    (see http://eigen.tuxfamily.org/)
    There you can access the column of a matrix A by saying A.col(i).
    And, using the ‘<<’ operator, you can build a matrix of out other
    matrices.
    See http://eigen.tuxfamily.org/dox-devel/QuickRefPage.html

    Helmut.

  3. admin says:

    - Hide quoted text — Show quoted text -

    On Fri, 21 Jan 2011 09:10:57 -0800, aruzinsky wrote:
    > 1. The license, at
    > http://eigen.tuxfamily.org/index.php?title=Licensing_FAQ, is overly
    > verbose bullshit.  The authors are assholes for even expecting
    > non-lawyers to read that.

    > 2. Those authors’ C++ classes are likely to conflict with one’s own C+ +
    > classes.  Amazingly those authors expect the World to revolve around
    > their C++ classes.

    > 3. In stark contrast, free, well written C code is available at
    > http://mymathlib.webtrellis.net/matrices/ which has no legal strings
    > other than copyright.  The developer must be a true philanthropist
    > because he or she does not even give their full name, only initials.

    > 4. The C code at http://mymathlib.webtrellis.net/matrices/ is easily
    > integrated into one’s own C++ classes.

    This shadows some light on the author!

    For all others:

    I have been using Eigen for teaching as well as for time critical
    applications with the best experience only.

    Anybody criticizing Eigen should show me a package which integrates well
    with C++ (no C-like packages, please) which has optional index checking
    with only 2 characters to change AND with a performance which even comes
    close to that of Eigen!

    But, believe whom you like to believe.

    Helmut Jarausch
    RWTH-Aachen University
    Germany

  4. admin says:

    "usuiisu" <ikuy…@gmail.com> wrote in message news:84788d70-de48-41ad-928b-cce21ffbb6cf@v26g2000yqf.googlegroups.com…

    | Hi,
    |
    | Can you think of some "selection matrix" to select only certain colum
    | vectors of a matrix?
    | For example,
    | X=[x1 x2 x3 x4] where xi is a column vector.
    | I would like to have a matrix J such that
    | XJ or JX will be [x1 x3] by selecting only 1st and 3rd columns of the
    | matrix.

    In PL/I —
        To get the first column , use x(*,1).
        Toget the third column, use x(*,3).

    The complete operation would be:
    y(*,1) = x(*,1);
    y(*,2) = x(*,3);

  5. admin says:

    "aruzinsky" <aruzin…@general-cathexis.com> wrote in message

    news:4c189118-bebc-4dbd-8ce6-f1d778c45d21@k22g2000yqh.googlegroups.com…

    >Please, stop perpetuating the delusion that the OP’s question has
    >something to do with some programming language.

    The OP’s question implies a programming approach.
    But even if that was not his intention, he can easily ignore my post.

    On the other hand, your own posts are about programming.

    On Jan 24, 5:01 am, "robin" <robi…@dodo.mapson.com.au> wrote:

    - Hide quoted text — Show quoted text -

    > "usuiisu" <ikuy…@gmail.com> wrote in
    > messagenews:84788d70-de48-41ad-928b-cce21ffbb6cf@v26g2000yqf.googlegroups.com…

    > | Hi,
    > |
    > | Can you think of some "selection matrix" to select only certain colum
    > | vectors of a matrix?
    > | For example,
    > | X=[x1 x2 x3 x4] where xi is a column vector.
    > | I would like to have a matrix J such that
    > | XJ or JX will be [x1 x3] by selecting only 1st and 3rd columns of the
    > | matrix.

    > In PL/I —
    > To get the first column , use x(*,1).
    > Toget the third column, use x(*,3).

    > The complete operation would be:
    > y(*,1) = x(*,1);
    > y(*,2) = x(*,3);

  6. admin says:

    "aruzinsky" <aruzin…@general-cathexis.com> wrote in message

    news:aeb4d8ef-3320-45bb-991a-dc1d17186ef8@l22g2000vbp.googlegroups.com…

    >On Jan 25, 7:35 am, "robin" <robi…@dodo.mapson.com.au> wrote:
    >> "aruzinsky" <aruzin…@general-cathexis.com> wrote in message

    >> news:4c189118-bebc-4dbd-8ce6-f1d778c45d21@k22g2000yqh.googlegroups.com…

    >> >Please, stop perpetuating the delusion that the OP’s question has
    >> >something to do with some programming language.

    >> The OP’s question implies a programming approach.
    >You are wrong.

    That’s up to the OP to decide, not you.
    I say that the question implies a programming approach.

    >> But even if that was not his intention, he can easily ignore my post.
    >But, you are wrong

    That’s for the OP to decide, not you.

    > and I am doing you a favor

    You are not doing anyone any favor.
    If you don’t like my post, you don’t have to read it.

Place your comment

You must be logged in to post a comment.