Preprocessing 0.2.0 normalizer

Lodestar.Preprocessing 0.2.0. This page is frozen at that release. Read the current documentation for what main says now. A link to a decision or a migration page follows main, and leaves the archive.

Home › Feature transforming

Normalizer

Scales each row to unit norm, at sklearn.preprocessing.Normalizer parity.

The one member of this package that fits nothing: every row is scaled by its own norm, so there is no statistic to learn from a training set and no state for an instance to carry. The reference is a transformer all the same, with a fit that only validates — so this is static.

Rows, not features. Every other member here scales a column; this one scales a row, which is what a distance or a dot product between two rows reads afterwards.

Members

Member	What it does
Normalizer.Transform	Scales each row of a matrix to unit norm.