def test_dldtheta(self):
a = 1.0 * rand()
b = a
while b <= a:
b = 10.0 * rand()
mu = rand() + 1.0
s = 10.0 * rand() + 1.0
p = Distributions.TruncatedExponentialPower({
'a': a,
'b': b,
'p': mu,
's': s
})
p.primary = ['p', 's']
dat = p.sample(100)
def f(arr):
p.array2primary(arr)
return np.sum(p.loglik(dat))
def df(arr):
p.array2primary(arr)
return np.sum(p.dldtheta(dat), axis=1)
arr0 = p.primary2array()
arr0 = abs(np.random.randn(len(arr0)))
err = optimize.check_grad(f, df, arr0)
print "Error in gradient: ", err
self.assertTrue(
err < 1e-02,
'Gradient error %.4g is greater than %.4g' % (err, 1e-02))
def test_loglik(self):
p1 = Distributions.TruncatedExponentialPower({
'a': -1.0,
'b': 2.0,
'p': 1.0,
's': 2.0
})
p2 = Distributions.TruncatedExponentialPower({
'a': -1.0,
'b': 2.0,
'p': 1.5,
's': 2.0
})
nsamples = 1000000
data = p2.sample(nsamples)
logZ = logsumexp(p1.loglik(data) - p2.loglik(data) - np.log(nsamples))
print "Estimated partition function: ", np.exp(logZ)
self.assertTrue(
np.abs(np.exp(logZ) - 1.0) < 0.1 * self.TolParam,
'Difference in estimated partition function (1.0) greater than' +
str(0.1 * self.TolParam))
def test_cdf(self):
print "Testing consistency of cdf and ppf"
sys.stdout.flush()
a = 1.0 * rand()
b = a
while b <= a:
b = 10.0 * rand()
mu = 3 * rand()
s = 2.0 * rand() + 1.0
p = Distributions.TruncatedExponentialPower({
'a': a,
'b': b,
'p': mu,
's': s
})
u = rand(10)
u2 = p.cdf(p.ppf(u))
self.assertFalse(
np.sum(np.abs(u - u2)) > self.TolParam,
'Difference u - cdf(ppf(u)) greater than %.4g' % (self.Tol, ))