normalizerow

compas.numerical.normalizerow(A, do_nan_to_num=True)[source]

Normalise the rows of matrix A.

Parameters:

Aarray: Matrix A represented as an (m x n) array.
do_nan_to_numbool: Convert NaNs and INF to numbers, default=True.

Returns:

array: Matrix of normalized row vectors (m x n).

Notes

Normalises the row vectors of A by the normrows, i.e. creates an array of vectors where the row vectors have length of unity.

Tiling is not necessary, because of NumPy’s broadcasting behaviour.

Examples

>>> normalizerow(array([[2, -1, 3,], [1, 0, 1], [0, 2, -1]]))
array([[ 0.53452248, -0.26726124,  0.80178373],
       [ 0.70710678,  0.        ,  0.70710678],
       [ 0.        ,  0.89442719, -0.4472136 ]])