11.3.1.4. astroML.linear_model.NadarayaWatson¶
-
class
astroML.linear_model.
NadarayaWatson
(kernel='gaussian', h=None, **kwargs)[source]¶ Nadaraya-Watson Kernel Regression
This is basically a gaussian-weighted moving average of points
- Parameters
- kernelstring
kernel is either “gaussian”, or one of the kernels available in sklearn.metrics.pairwise.
- hfloat or array_like
width of kernel. If array, its length must be the number of dimensions in the training data
- Additional keyword arguments are passed to the kernel.
Methods
fit
predict