def set_data(self, data):
X, Y = data
self._transform = astransform(X)
self._X = X
self._is_transform = id(self._X) == id(
self._transform) # i.e. astransform was a nullop
self._Y = Y
def restricted_Mest(Mest_loss, active, solve_args={'min_its':50, 'tol':1.e-10}):
"""
Fit a restricted model using only columns `active`.
Parameters
----------
Mest_loss : objective function
A GLM loss.
active : ndarray
Which columns to use.
solve_args : dict
Passed to `solve`.
Returns
-------
soln : ndarray
Solution to restricted problem.
"""
X, Y = Mest_loss.data
if not Mest_loss._is_transform and hasattr(Mest_loss, 'saturated_loss'): # M_est is a glm
X_restricted = X[:,active]
loss_restricted = rr.affine_smooth(Mest_loss.saturated_loss, X_restricted)
else:
I_restricted = ra.selector(active, ra.astransform(X).input_shape[0], ra.identity((active.sum(),)))
loss_restricted = rr.affine_smooth(Mest_loss, I_restricted.T)
beta_E = loss_restricted.solve(**solve_args)
return beta_E
def test_admm(n=100, p=10):
X = np.random.standard_normal((n, p))
Y = np.random.standard_normal(n)
loss = rr.squared_error(X, Y)
D = np.identity(p)
pen = rr.l1norm(p, lagrange=1.5)
ADMM = admm_problem(loss, pen, ra.astransform(D), 0.5)
ADMM.solve(niter=1000)
coef1 = ADMM.atom_coefs
problem2 = rr.simple_problem(loss, pen)
coef2 = problem2.solve(tol=1.e-12, min_its=500)
np.testing.assert_allclose(coef1, coef2, rtol=1.e-3, atol=1.e-4)