def testInitPolynomialModel(self):
# Defining a polynomial model with no covariance matrix should be the
# same as one with a diagonal covariance matrix with ones on the diagonal.
orderList = [0,2]
nParameters = len(orderList)
covariateName = "t"
regressorNames = ["1", "t^2"]
polyModel1 = PolynomialModel(self.time, covariateName, orderList)
polyModel2 = PolynomialModel(self.time, covariateName, orderList, np.diag(np.ones(self.nObservations)))
self.assertTrue(np.alltrue(polyModel1.designMatrix() == polyModel2.designMatrix()))
def testInitPolynomialModel(self):
# Define the input, and derive what the output should be
orderList = [0,2]
nParameters = len(orderList)
covariateName = "t"
regressorNames = ["1", "t^2"]
designMatrix = np.empty((self.nObservations, nParameters))
for n in range(len(orderList)):
designMatrix[:,n] = self.time**orderList[n]
# Assert if the output is what it should be
polyModel = PolynomialModel(self.time, covariateName, orderList)
self.assertTrue(polyModel.nParameters() == nParameters)
self.assertTrue(polyModel.regressorNames() == regressorNames)
self.assertTrue(polyModel.nObservations() == self.nObservations)
self.assertTrue(np.alltrue(polyModel.designMatrix() == designMatrix))
