scipy.special.boxcox¶
- scipy.special.boxcox(x, lmbda) = <ufunc 'boxcox'>¶
Compute the Box-Cox transformation.
The Box-Cox transformation is:
y = (x**lmbda - 1) / lmbda if lmbda != 0 log(x) if lmbda == 0
Returns nan if x < 0. Returns -inf if x == 0 and lmbda < 0.
New in version 0.14.0.
Parameters: x : array_like
Data to be transformed.
lmbda : array_like
Power parameter of the Box-Cox transform.
Returns: y : array
Transformed data.
Examples
>>> boxcox([1, 4, 10], 2.5) array([ 0. , 12.4 , 126.09110641]) >>> boxcox(2, [0, 1, 2]) array([ 0.69314718, 1. , 1.5 ])