def local_error_test(self):
"""Test MeanSquaredError.local_error."""
localErrors = [0.04, 0.09, 0.16, 0.25, 0.25, 0.09, 0.04, 0.36, 0.09, 0.25, 0]
mse = MeanSquaredError()
for i in range(len(self.dataOrg)):
calc_local_error = mse.local_error([self.dataOrg[i]], [self.dataCalc[i]])
self.assertEquals("%.3f" % calc_local_error,"%.3f" % localErrors[i])
def local_error_test(self):
"""Test MeanSquaredError.local_error."""
localErrors = [0.04, 0.09, 0.16, 0.25, 0.25, 0.09, 0.04, 0.36, 0.09, 0.25, 0]
mse = MeanSquaredError()
for i in xrange(len(self.dataOrg)):
calc_local_error = mse.local_error([self.dataOrg[i]], [self.dataCalc[i]])
self.assertEquals("%.3f" % calc_local_error,"%.3f" % localErrors[i])
def error_calculation_test(self):
"""Test the calculation of the MeanSquaredError."""
tsOrg = TimeSeries()
tsCalc = TimeSeries()
for idx in range(len(self.dataOrg)):
tsOrg.add_entry(float(idx), self.dataOrg[idx])
tsCalc.add_entry(float(idx), self.dataCalc[idx])
mse = MeanSquaredError()
mse.initialize(tsOrg, tsCalc)
self.assertEquals("0.1472", str(mse.get_error())[:6])
def error_calculation_test(self):
"""Test the calculation of the MeanSquaredError."""
tsOrg = TimeSeries()
tsCalc = TimeSeries()
for idx in xrange(len(self.dataOrg)):
tsOrg.add_entry(float(idx), self.dataOrg[idx])
tsCalc.add_entry(float(idx), self.dataCalc[idx])
mse = MeanSquaredError()
mse.initialize(tsOrg, tsCalc)
self.assertEquals("0.1472", str(mse.get_error())[:6])
def error_calculation_test(self):
"""Test for a valid error calculation."""
dataOrg = [[0,0], [1,1], [2,2], [3,3], [4,4], [5,5], [6, 6], [7,7], [8,8], [9,9]]
dataCalc = [[0,1], [1,3], [2,5], [3,0], [4,3], [5,5], [6.1,6], [7,3], [8.1,8], [9,8]]
# difference: 1 2 3 3 1 0 NA 4 NA 1
# local errors: 1 4 9 9 1 0 NA 16 NA 1
tsOrg = TimeSeries.from_twodim_list(dataOrg)
tsCalc = TimeSeries.from_twodim_list(dataCalc)
tsOrg = TimeSeries.from_twodim_list(dataOrg)
tsCalc = TimeSeries.from_twodim_list(dataCalc)
mse = MeanSquaredError(80.0)
mse.initialize(tsOrg, tsCalc)
assert str(mse.get_error()) == "5.125"
def number_of_comparisons_test(self):
""" Test BaseErrorMeasure.initialize for behaviour if not enough dates match."""
dataOrg = [[0,0],[1,1],[2,2],[3,3],[4,4],[5,5],[6,6],[7,7],[8,8],[9,9]]
dataCalc = [[0,0],[1,1],[2,2],[3,3],[4,4],[5.1,5],[6.1,6],[7.1,7],[8.1,8],[9.1,9]]
tsOrg = TimeSeries.from_twodim_list(dataOrg)
tsCalc = TimeSeries.from_twodim_list(dataCalc)
bem = BaseErrorMeasure(60.0)
#prevent NotImplementedError
bem.local_error = lambda a,b: 1
mse = MeanSquaredError(80.0)
# only 50% of the original TimeSeries have a corresponding partner
if mse.initialize(tsOrg, tsCalc):
assert False # pragma: no cover
if not mse.initialize(tsOrg, tsOrg):
assert False # pragma: no cover
