From Eigenvector Research Documentation Wiki
Jump to navigation Jump to search


Trilinear decomposition.


model = tld(x,ncomp,scl,plots)


The trilinear decomposition can be used to decompose a 3-way array as the summation over the outer product of triads of vectors. Inputs are the 3 way array x and the number of components to estimate ncomp. Optional input variables include scales for each of of the array axes, (scl1, scl2, scl3). These axes can be entered as 0 or [] placeholders. The output of TLD is a structured array (model) containing all of the model elements in the following fields:

  • date: model creation date stamp
  • time: model creation time stamp
  • size: size of the original input array
  • loads: 1 by 3 cell array of the loadings in each dimension
  • res: 1 by 3 cell array residuals summed over each dimension
  • scl: 1 by 3 cell array with scales for plotting loads

Note that the model loadings are presented as unit vectors for the first two dimensions, remaining scale information is incorporated into the final (third) dimension.

See Also

gram, outerm, parafac