segreg.model.OneBkptSegRegEstimator.params

property OneBkptSegRegEstimator.params

Returns the fitted parameters.

The returned parameters are, in order,

[u, v, m1, m2, sigma]

More generally, if the (fitted) model is of the form:

\[y_i = f(x_i; \theta) + \varepsilon_i\]

where \(f(x)\) is a function of x depending on the parameter vector \(\theta\), and \(\{\varepsilon_i\}\) is i.i.d. random error of mean zero and variance \(\sigma^2\), then the returned parameter vector is of the form:

\[[\theta_1, \dots, \theta_k, \sigma]\]

where \(\theta = [\theta_1, \dots, \theta_k]\).

Notes

The error variance \(\sigma^2\) is often considered a nuisance parameter, and is not required for many applications.

Raises

Exception – If fit has not been called.

Returns

params

Return type

array-like