def evaluate(X_tr, y_tr, X_te, y_te, T):
grid = sg.RegularGridConfiguration()
grid.dim_ = 10
grid.level_ = 4
grid.t_ = T
grid.type_ = sg.GridType_ModLinear
adapt = sg.AdaptivityConfiguration()
adapt.numRefinements_ = 5
adapt.noPoints_ = 3
solv = sg.SLESolverConfiguration()
solv.maxIterations_ = 50
solv.eps_ = 1e-5
solv.threshold_ = 1e-5
solv.type_ = sg.SLESolverType_CG
final_solv = solv
final_solv.maxIterations = 200
regular = sg.RegularizationConfiguration()
regular.type_ = sg.RegularizationType_Identity
regular.exponentBase_ = 1.0
regular.lambda_ = 1e-3
## Create the estimator, train it with the training data and then return the error
## for the testing set.
estimator = sg.RegressionLearner(grid, adapt, solv, final_solv, regular)
estimator.train(X_tr, y_tr)
print(estimator.getGridSize())
return estimator.getMSE(X_te, y_te)
def make_estimator(lambda_reg, prior):
grid = sg.RegularGridConfiguration()
grid.dim_ = 4
grid.level_ = 5
grid.type_ = sg.GridType_ModLinear
adapt = sg.AdaptivityConfiguration()
adapt.numRefinements_ = 5
adapt.noPoints_ = 3
solv = sg.SLESolverConfiguration()
solv.maxIterations_ = 50
solv.eps_ = 10e-6
solv.threshold_ = 10e-6
solv.type_ = sg.SLESolverType_CG
final_solv = solv
final_solv.maxIterations = 200
regular = sg.RegularizationConfiguration()
regular.type_ = sg.RegularizationType_Diagonal
regular.exponentBase_ = prior
regular.lambda_ = lambda_reg
estimator = sg.RegressionLearner(grid, adapt, solv, final_solv, regular)
return estimator
def make_estimator(penalty, l1_ratio, lambda_reg):
grid = sg.RegularGridConfiguration()
grid.dim_ = 10
grid.level_ = 2
grid.type_ = sg.GridType_ModLinear
adapt = sg.AdaptivityConfiguration()
adapt.numRefinements_ = 0
adapt.noPoints_ = 0
solv = sg.SLESolverConfiguration()
solv.maxIterations_ = 500
solv.threshold_ = 10e-10
solv.type_ = sg.SLESolverType_FISTA
final_solv = solv
regular = sg.RegularizationConfiguration()
regular.type_ = penalty
regular.exponentBase_ = 1.0
regular.lambda_ = lambda_reg
regular.l1_ratio_ = l1_ratio
estimator = sg.RegressionLearner(grid, adapt, solv, final_solv,regular)
return estimator