Matrix Multiplication Filter
Performs a matrix multiplication with a given training-only set of basis vectors.
Category |
|
Node |
|
Parameters |
(None) |
Inputs |
Input: the input data Matrix: the basis vectors to multiply with |
Outputs |
Counts: the number of peaks |
Effect of the Filter
Given a set of input data that correspond to coefficients in a specific basis, and a training-only matrix that contains the basis vectors of that basis, perform a matrix multiplication to calculate the input in the original basis.
This may be used in conjunction with the Subspace Projection Filter or the Abundance Determination Filter.
The following screenshots show how this filter may be used within a graph:
This filter is most useful if the coefficients (e.g. the abundances) have been modified. For example, it is possible to use a Clamp Values Filter to clamp the coefficients to a specific value range and then perform the matrix multiplication to obtain the input data projected in this manner in the original vector space.