def testInverseCumulativeNormalPickle(self):
benchmark_norm = InverseCumulativeNormal(average, sigma)
f = tempfile.NamedTemporaryFile('w+b', delete=False)
pickle.dump(benchmark_norm, f)
f.close()
with open(f.name, 'rb') as f2:
pickled_norm = pickle.load(f2)
self.assertEqual(benchmark_norm, pickled_norm)
def testNormal(self):
invCumStandardNormal = InverseCumulativeNormal()
check = invCumStandardNormal(0.5)
self.assertAlmostEqual(
check, 0.0, 10,
"inverse cumulative of the standard normal at 0.5 is {0:f}"
"\n instead of zero: something is wrong!".format(check))
normal = NormalDistribution(average, sigma)
cum = CumulativeNormalDistribution(average, sigma)
invCum = InverseCumulativeNormal(average, sigma)
invCumAcc = InverseCumulativeNormal(average, sigma, fullAccuracy=True)
numberOfStandardDeviation = 6
xMin = average - numberOfStandardDeviation * sigma
xMax = average + numberOfStandardDeviation * sigma
N = 100001
h = (xMax - xMin) / (N - 1)
x = [xMin + i * h for i in range(N)]
y = [gaussian(v) for v in x]
yd = [gaussianDerivative(v) for v in x]
temp = [normal(v) for v in x]
for i, (expected, calculated) in enumerate(zip(y, temp)):
self.assertAlmostEqual(
expected, calculated, 15, "at index {0:d}\n"
"Expected: {1:f}\n"
"Calculated: {2:f}".format(i, expected, calculated))
temp = [cum(v) for v in x]
temp = [invCum(v) for v in temp]
for i, (expected, calculated) in enumerate(zip(x, temp)):
self.assertAlmostEqual(
expected, calculated, 7, "at index {0:d}\n"
"Expected gaussian: {1:f}\n"
"Calculated Gaussian: {2:f}".format(i, expected, calculated))
temp = [cum(v) for v in x]
temp = [invCumAcc(v) for v in temp]
for i, (expected, calculated) in enumerate(zip(x, temp)):
self.assertAlmostEqual(
expected, calculated, 7, "at index {0:d}\n"
"Expected gaussian: {1:.9f}\n"
"Calculated Gaussian: {2:.9f}".format(i, expected, calculated))
temp = [cum.derivative(v) for v in x]
for i, (expected, calculated) in enumerate(zip(y, temp)):
self.assertAlmostEqual(
expected, calculated, 15, "at index {0:d}\n"
"Expected: {1:f}\n"
"Calculated: {2:f}".format(i, expected, calculated))
temp = [normal.derivative(v) for v in x]
for i, (expected, calculated) in enumerate(zip(yd, temp)):
self.assertAlmostEqual(
expected, calculated, 15, "at index {0:d}\n"
"Expected: {1:f}\n"
"Calculated: {2:f}".format(i, expected, calculated))
# test nan value returning
self.assertTrue(math.isnan(invCum(-0.5)))
def testInverseCumulativeNormal(self):
norm = InverseCumulativeNormal(average, sigma, True)
copied = copy.deepcopy(norm)
self.assertEqual(norm, copied)