# Least Mean Squares Offline

The offline LMS rule a non-iterative method for computing a weight matrix for associating input vectors and target vectors in a training dataset (see training dialog).

The lms offline trainer is configured using a special panel, shown here, and is used primarily used by the least mean square network and echo state network, though it can also be used by scripts and in other ways. The following parameter can be modified.

Note for this algorithm you have to make sure the output node range is greater than the range of the target values. Also it sometimes it helps to change output nodes to linear.

Training Type: Wiener-Hopf: The wiener-hopf method for lms offline.

Training Type: Moore-Penrose: The moore penrose method for lms offline. Moore-Penrose is stable in some cases when Wiener Hopf is not.

Ridge regression, Noise: Ridge regression and adding noise help when a matrix that is rank deficient. These options help prevent over-fitting by making rank slightly higher.