Stats on polar vectors

From: Scott Seidman (namdiesttocs_at_mindspring.com)
Date: 01/06/05

  • Next message: Ners: "Re: Snatching Residuals" 
    Date: 6 Jan 2005 19:57:11 GMT

    I'm using multivariate techniques to compare vectors, after converting the
    vectors from a polar to a cartesian coordinate schemes. What I REALLY want
    to do is compare phase and magnitude separately, but I'm having some
    problems for a variety of reasons. The first is that if the magnitude is
    small, that is the origin of the cartesian coordinate system is in the
    distribution, phase is pretty much indeterminate. The next is that phase
    is actually periodic in 360 degrees. So, in many ways things make more
    "sense" when comparing these multivariate distributions in cartesian space,
    but not so in polar space.

    One thing I've been thinking about is just calculating the vector
    difference between sets in cartesian space, and if the result has a
    meaningful phase, stating the significant phase difference as the result of
    this subtraction. I'd imagine it would be OK to state that the phase
    difference is non-zero so long as the y-component CI does not cross the x-
    axis.

    Am I moving in the right direction?

    Thanks
    Scott

  • Next message: Ners: "Re: Snatching Residuals"

    Relevant Pages