regression

Module Contents

Classes

CCS_regressor

class regression.CCS_regressor(N=100, xmin=0, xmax=1, dx=None, sw=False, eps=False)
merge_intervals(x)
rubber_band()
static solver(pp, qq, gg, hh, aa, bb, maxiter=300, tol=(1e-10, 1e-10, 1e-10))

The solver for the objective.

Parameters:

  • pp (matrix) – P matrix as per standard Quadratic Programming(QP) notation.

  • qq (matrix) – q matrix as per standard QP notation.

  • gg (matrix) – G matrix as per standard QP notation.

  • hh (matrix) – h matrix as per standard QP notation

  • aa (matrix) – A matrix as per standard QP notation.

  • bb (matrix) – b matrix as per standard QP notation

  • maxiter (int, optional) – maximum iteration steps (default: 300).

  • tol (tuple, optional) – tolerance value of the solution (default: (1e-10, 1e-10, 1e-10)).

Returns:

dictionary containing solution details

Return type:

sol (dict)

fit(x, y)
const(n_switch)
predict(x)
spline_construction(N)

This function constructs the matrices A, B, C, D.

Parameters:

N – Number of knots

Returns:

constructed matrices

Return type:

cc, dd, bb, aa (matrices)

model(x)

Constructs the v matrix.

Parameters:

self –

Returns:

The v matrix for a pair.

Return type:

ndarray