Purpose

---

  This function computes the covariance matrix for a set of
  multidimensional data. It optionally returns the principal axes, of
  the distribution, along with the variance along each principal
  axis.
  The principal axes and variances describe the orientation and size
  of the "error ellipse" for multivariate gaussians.

Input

---

  data: An (ndimension) x (n point) data array.

Keywords

---

  paxis: On output, will hold the principal axes of the distribution.
  An (ndim x ndim) array. The ith row of this array lists the ith
  principal axis.
  pvar: On output, will hold the variance along each principal axis.
  An (ndim) vector.
  mean: On output, will hold the mean of the distribution.
  weights: An option (npoint) array that will weight each data point

Output

---

  The covariance array. An (ndimension) x (ndimension) array.

Modification History

---

  March 2010: Written by Chris Beaumont
  December 2010: Added paxis, pvar, and mean keywords. cnb.