models.Gam
A generalized additive model: a Glm plus P-spline smooths of
Usage
models.Gam()numeric design columns, with smoothing chosen by GCV or UBRE.
Parameters
glm: Glm-
Family, link and dispersion.
smooths: list of str or (str, int)-
The design columns to smooth, optionally with the number of basis functions (10 by default).
smoothing: str or list of float = "auto"-
"auto"(UBRE for a fixed dispersion, GCV otherwise),"gcv","ubre", or fixed smoothing parameters, one per smooth.
Examples
>>> import math
>>> from prospicio.models import Design, Gam, Glm
>>> x = [i / 99 for i in range(100)]
>>> y = [math.sin(6 * v) for v in x]
>>> d = Design([[1.0] * 100, x], ["(Intercept)", "x"])
>>> fit = Gam(Glm("gaussian"), ["x"]).fit(d, y)
>>> abs(fit.predict(d)[50] - y[50]) < 0.01True
Methods
| Name | Description |
|---|---|
| fit() | Fits the model. |
fit()
Fits the model.
Usage
fit(design, y)Parameters
design: Design-
Includes the raw columns to smooth.
y: list of float
Returns
GamFit