def test_fit(self):
""" optimization test """
self.vc.optimize(verbose=False)
params = self.vc.getScales()
if self.generate:
self.D['params_true'] = params
data.dump(self.D,self.dataset)
self.generate=False
params_true = self.D['params_true']
RV = ((SP.absolute(params)-SP.absolute(params_true))**2).max()
self.assertTrue(RV<1e-6)
def test_fit(self):
""" optimization test """
self.vc.optimize(verbose=False)
params = self.vc.getScales()
if self.generate:
self.D['params_true'] = params
data.dump(self.D, self.dataset)
self.generate = False
params_true = self.D['params_true']
RV = ((SP.absolute(params) - SP.absolute(params_true))**2).max()
self.assertTrue(RV < 1e-6)
def test_fit(self):
""" optimization test """
self.vc.optimize(verbose=False)
params = self.vc.getScales()
if self.generate:
self.D['params_true'] = params
data.dump(self.D,self.dataset)
self.generate=False
params_true = self.D['params_true']
RV = ((SP.absolute(params)-SP.absolute(params_true))**2).max()
#permit more flexibility, as we set a few values to NAN
self.assertTrue(RV<1e-4)
def test_fit(self):
""" optimization test """
self.vc.optimize(verbose=False)
params = self.vc.getScales()
if self.generate:
self.D['params_true'] = params
data.dump(self.D, self.dataset)
self.generate = False
params_true = self.D['params_true']
RV = ((SP.absolute(params) - SP.absolute(params_true))**2).max()
#permit more flexibility, as we set a few values to NAN
self.assertTrue(RV < 1e-4)
def test_fitFast(self):
""" optimization test """
self.vc.optimize(fast=True,verbose=False)
params = self.vc.getScales()
if self.generate:
self.D['params_true'] = params
data.dump(self.D,self.dataset)
self.generate=False
params_true = self.D['params_true']
#make sign invariant
RV = ((SP.absolute(params)-SP.absolute(params_true))**2).max()<1e-6
self.assertTrue(RV)
def test_fitFast(self):
""" optimization test """
self.vc.optimize(fast=True, verbose=False)
params = self.vc.getScales()
if self.generate:
self.D['params_true'] = params
data.dump(self.D, self.dataset)
self.generate = False
params_true = self.D['params_true']
#make sign invariant
RV = ((SP.absolute(params) - SP.absolute(params_true))**2).max() < 1e-6
self.assertTrue(RV)
def test_fit(self):
""" test fitting """
self.lmmlasso.set_params(alpha=1e-1)
self.lmmlasso.fit(self.D['X'], self.D['y'], self.D['K'])
params = self.lmmlasso.coef_
yhat = self.lmmlasso.predict(self.D['X'], self.D['K'])
if self.generate:
self.D['params_true'] = params
self.D['yhat'] = yhat
data.dump(self.D, self.dataset)
self.generate = False
params_true = self.D['params_true']
yhat_true = self.D['yhat']
RV = ((SP.absolute(params) - SP.absolute(params_true))**2).max()
np.testing.assert_almost_equal(RV, 0., decimal=4)
RV = ((SP.absolute(yhat) - SP.absolute(yhat_true))**2).max()
np.testing.assert_almost_equal(RV, 0., decimal=2)
def test_fit(self):
""" test fitting """
self.lmmlasso.set_params(alpha=1e-1)
self.lmmlasso.fit(self.D['X'],self.D['y'],self.D['K'])
params = self.lmmlasso.coef_
yhat = self.lmmlasso.predict(self.D['X'],self.D['K'])
if self.generate:
self.D['params_true'] = params
self.D['yhat'] = yhat
data.dump(self.D,self.dataset)
self.generate=False
params_true = self.D['params_true']
yhat_true = self.D['yhat']
RV = ((SP.absolute(params)-SP.absolute(params_true))**2).max()
np.testing.assert_almost_equal(RV, 0., decimal=4)
RV = ((SP.absolute(yhat)-SP.absolute(yhat_true))**2).max()
np.testing.assert_almost_equal(RV, 0., decimal=2)