Hat Matrix#
A hat matrix is an object used in kernel smoothing (KernelSmoother
) and
kernel regression (KernelRegression
) algorithms.
Those algorithms estimate the desired values as a weighted mean of train data. The different Hat matrix types define how these weights are calculated.
See the links below for more information.
|
Hat Matrix. |
Nadaraya-Watson method. |
|
Local linear regression method. |
|
K-nearest neighbour kernel method. |