def test_binomial_numeric():
nvals = range(5)
pvals = [0, Rational(1, 4), S.Half, Rational(3, 4), 1]
for n in nvals:
for p in pvals:
X = Binomial('X', n, p)
assert E(X) == n * p
assert variance(X) == n * p * (1 - p)
if n > 0 and 0 < p < 1:
assert skewness(X) == (1 - 2 * p) / sqrt(n * p * (1 - p))
assert kurtosis(X) == 3 + (1 - 6 * p * (1 - p)) / (n * p *
(1 - p))
for k in range(n + 1):
assert P(Eq(X, k)) == binomial(n, k) * p**k * (1 - p)**(n - k)
def test_weibull():
a, b = symbols('a b', positive=True)
# FIXME: simplify(E(X)) seems to hang without extended_positive=True
# On a Linux machine this had a rapid memory leak...
# a, b = symbols('a b', positive=True)
X = Weibull('x', a, b)
assert E(X).expand() == a * gamma(1 + 1 / b)
assert variance(X).expand() == (a**2 * gamma(1 + 2 / b) - E(X)**2).expand()
assert simplify(skewness(X)) == (
2 * gamma(1 + 1 / b)**3 - 3 * gamma(1 + 1 / b) * gamma(1 + 2 / b) +
gamma(1 + 3 / b)) / (-gamma(1 + 1 / b)**2 +
gamma(1 + 2 / b))**Rational(3, 2)
assert simplify(kurtosis(X)) == (-3*gamma(1 + 1/b)**4 +\
6*gamma(1 + 1/b)**2*gamma(1 + 2/b) - 4*gamma(1 + 1/b)*gamma(1 + 3/b) + gamma(1 + 4/b))/(gamma(1 + 1/b)**2 - gamma(1 + 2/b))**2
def test_binomial_symbolic():
n = 2
p = symbols("p", positive=True)
X = Binomial("X", n, p)
t = Symbol("t")
assert simplify(E(X)) == n * p == simplify(moment(X, 1))
assert simplify(variance(X)) == n * p * (1 - p) == simplify(cmoment(X, 2))
assert cancel((skewness(X) - (1 - 2 * p) / sqrt(n * p * (1 - p)))) == 0
assert cancel((kurtosis(X)) - (3 + (1 - 6 * p * (1 - p)) / (n * p *
(1 - p)))) == 0
assert (characteristic_function(X)(t) == p**2 * exp(2 * I * t) + 2 * p *
(-p + 1) * exp(I * t) + (-p + 1)**2)
assert (moment_generating_function(X)(t) == p**2 * exp(2 * t) + 2 * p *
(-p + 1) * exp(t) + (-p + 1)**2)
# Test ability to change success/failure winnings
H, T = symbols("H T")
Y = Binomial("Y", n, p, succ=H, fail=T)
assert simplify(E(Y) - (n * (H * p + T * (1 - p)))) == 0
# test symbolic dimensions
n = symbols("n")
B = Binomial("B", n, p)
raises(NotImplementedError, lambda: P(B > 2))
assert density(B).dict == Density(BinomialDistribution(n, p, 1, 0))
assert set(density(B).dict.subs(n, 4).doit().keys()) == set(
[S.Zero, S.One, S(2), S(3), S(4)])
assert set(density(B).dict.subs(n, 4).doit().values()) == set([
(1 - p)**4,
4 * p * (1 - p)**3,
6 * p**2 * (1 - p)**2,
4 * p**3 * (1 - p),
p**4,
])
k = Dummy("k", integer=True)
assert E(B > 2).dummy_eq(
Sum(
Piecewise(
(
k * p**k * (1 - p)**(-k + n) * binomial(n, k),
(k >= 0) & (k <= n) & (k > 2),
),
(0, True),
),
(k, 0, n),
))
def test_binomial_symbolic():
n = 2 # Because we're using for loops, can't do symbolic n
p = symbols('p', positive=True)
X = Binomial('X', n, p)
t = Symbol('t')
assert simplify(E(X)) == n*p == simplify(moment(X, 1))
assert simplify(variance(X)) == n*p*(1 - p) == simplify(cmoment(X, 2))
assert cancel((skewness(X) - (1 - 2*p)/sqrt(n*p*(1 - p)))) == 0
assert cancel((kurtosis(X)) - (3 + (1 - 6*p*(1 - p))/(n*p*(1 - p)))) == 0
assert characteristic_function(X)(t) == p ** 2 * exp(2 * I * t) + 2 * p * (-p + 1) * exp(I * t) + (-p + 1) ** 2
assert moment_generating_function(X)(t) == p ** 2 * exp(2 * t) + 2 * p * (-p + 1) * exp(t) + (-p + 1) ** 2
# Test ability to change success/failure winnings
H, T = symbols('H T')
Y = Binomial('Y', n, p, succ=H, fail=T)
assert simplify(E(Y) - (n*(H*p + T*(1 - p)))) == 0
def test_gamma():
k = Symbol("k", positive=True)
theta = Symbol("theta", positive=True)
X = Gamma('x', k, theta)
assert density(X)(x) == x**(k - 1) * theta**(-k) * exp(
-x / theta) / gamma(k)
assert cdf(X, meijerg=True)(z) == Piecewise(
(-k * lowergamma(k, 0) / gamma(k + 1) +
k * lowergamma(k, z / theta) / gamma(k + 1), z >= 0), (0, True))
# assert simplify(variance(X)) == k*theta**2 # handled numerically below
assert E(X) == moment(X, 1)
k, theta = symbols('k theta', positive=True)
X = Gamma('x', k, theta)
assert E(X) == k * theta
assert variance(X) == k * theta**2
assert skewness(X).expand() == 2 / sqrt(k)
assert kurtosis(X).expand() == 3 + 6 / k
def test_Hermite():
a1 = Symbol("a1", positive=True)
a2 = Symbol("a2", negative=True)
raises(ValueError, lambda: Hermite("H", a1, a2))
a1 = Symbol("a1", negative=True)
a2 = Symbol("a2", positive=True)
raises(ValueError, lambda: Hermite("H", a1, a2))
a1 = Symbol("a1", positive=True)
x = Symbol("x")
H = Hermite("H", a1, a2)
assert moment_generating_function(H)(x) == exp(a1 * (exp(x) - 1) + a2 *
(exp(2 * x) - 1))
assert characteristic_function(H)(x) == exp(a1 * (exp(I * x) - 1) + a2 *
(exp(2 * I * x) - 1))
assert E(H) == a1 + 2 * a2
H = Hermite("H", a1=5, a2=4)
assert density(H)(2) == 33 * exp(-9) / 2
assert E(H) == 13
assert variance(H) == 21
assert kurtosis(H) == Rational(464, 147)
assert skewness(H) == 37 * sqrt(21) / 441
def test_robust_soliton():
raises(ValueError, lambda : RobustSoliton('robSol', -12, 0.1, 0.02))
raises(ValueError, lambda : RobustSoliton('robSol', 13, 1.89, 0.1))
raises(ValueError, lambda : RobustSoliton('robSol', 15, 0.6, -2.31))
f = Function('f')
raises(ValueError, lambda : density(RobustSoliton('robSol', 15, 0.6, 0.1)).pmf(f))
k = Symbol('k', integer=True, positive=True)
delta = Symbol('delta', positive=True)
c = Symbol('c', positive=True)
robSol = RobustSoliton('robSol', k, delta, c)
assert density(robSol).low == 1
assert density(robSol).high == k
k_vals = [10, 20, 50]
delta_vals = [0.2, 0.4, 0.6]
c_vals = [0.01, 0.03, 0.05]
for x in k_vals:
for y in delta_vals:
for z in c_vals:
assert E(robSol.subs({k: x, delta: y, c: z})) == moment(robSol.subs({k: x, delta: y, c: z}), 1)
assert variance(robSol.subs({k: x, delta: y, c: z})) == cmoment(robSol.subs({k: x, delta: y, c: z}), 2)
assert skewness(robSol.subs({k: x, delta: y, c: z})) == smoment(robSol.subs({k: x, delta: y, c: z}), 3)
assert kurtosis(robSol.subs({k: x, delta: y, c: z})) == smoment(robSol.subs({k: x, delta: y, c: z}), 4)
def test_dice():
# TODO: Make iid method!
X, Y, Z = Die('X', 6), Die('Y', 6), Die('Z', 6)
a, b, t, p = symbols('a b t p')
assert E(X) == 3 + S.Half
assert variance(X) == S(35) / 12
assert E(X + Y) == 7
assert E(X + X) == 7
assert E(a * X + b) == a * E(X) + b
assert variance(X + Y) == variance(X) + variance(Y) == cmoment(X + Y, 2)
assert variance(X + X) == 4 * variance(X) == cmoment(X + X, 2)
assert cmoment(X, 0) == 1
assert cmoment(4 * X, 3) == 64 * cmoment(X, 3)
assert covariance(X, Y) == S.Zero
assert covariance(X, X + Y) == variance(X)
assert density(Eq(cos(X * S.Pi), 1))[True] == S.Half
assert correlation(X, Y) == 0
assert correlation(X, Y) == correlation(Y, X)
assert smoment(X + Y, 3) == skewness(X + Y)
assert smoment(X + Y, 4) == kurtosis(X + Y)
assert smoment(X, 0) == 1
assert P(X > 3) == S.Half
assert P(2 * X > 6) == S.Half
assert P(X > Y) == S(5) / 12
assert P(Eq(X, Y)) == P(Eq(X, 1))
assert E(X, X > 3) == 5 == moment(X, 1, 0, X > 3)
assert E(X, Y > 3) == E(X) == moment(X, 1, 0, Y > 3)
assert E(X + Y, Eq(X, Y)) == E(2 * X)
assert moment(X, 0) == 1
assert moment(5 * X, 2) == 25 * moment(X, 2)
assert quantile(X)(p) == Piecewise((nan, (p > S.One) | (p < S(0))),\
(S.One, p <= S(1)/6), (S(2), p <= S(1)/3), (S(3), p <= S.Half),\
(S(4), p <= S(2)/3), (S(5), p <= S(5)/6), (S(6), p <= S.One))
assert P(X > 3, X > 3) == S.One
assert P(X > Y, Eq(Y, 6)) == S.Zero
assert P(Eq(X + Y, 12)) == S.One / 36
assert P(Eq(X + Y, 12), Eq(X, 6)) == S.One / 6
assert density(X + Y) == density(Y + Z) != density(X + X)
d = density(2 * X + Y**Z)
assert d[S(22)] == S.One / 108 and d[S(4100)] == S.One / 216 and S(
3130) not in d
assert pspace(X).domain.as_boolean() == Or(
*[Eq(X.symbol, i) for i in [1, 2, 3, 4, 5, 6]])
assert where(X > 3).set == FiniteSet(4, 5, 6)
assert characteristic_function(X)(t) == exp(6 * I * t) / 6 + exp(
5 * I * t) / 6 + exp(4 * I * t) / 6 + exp(3 * I * t) / 6 + exp(
2 * I * t) / 6 + exp(I * t) / 6
assert moment_generating_function(X)(
t) == exp(6 * t) / 6 + exp(5 * t) / 6 + exp(4 * t) / 6 + exp(
3 * t) / 6 + exp(2 * t) / 6 + exp(t) / 6
# Bayes test for die
BayesTest(X > 3, X + Y < 5)
BayesTest(Eq(X - Y, Z), Z > Y)
BayesTest(X > 3, X > 2)
# arg test for die
raises(ValueError, lambda: Die('X', -1)) # issue 8105: negative sides.
raises(ValueError, lambda: Die('X', 0))
raises(ValueError, lambda: Die('X', 1.5)) # issue 8103: non integer sides.
# symbolic test for die
n, k = symbols('n, k', positive=True)
D = Die('D', n)
dens = density(D).dict
assert dens == Density(DieDistribution(n))
assert set(dens.subs(n, 4).doit().keys()) == set([1, 2, 3, 4])
assert set(dens.subs(n, 4).doit().values()) == set([S(1) / 4])
k = Dummy('k', integer=True)
assert E(D).dummy_eq(Sum(Piecewise((k / n, k <= n), (0, True)), (k, 1, n)))
assert variance(D).subs(n, 6).doit() == S(35) / 12
ki = Dummy('ki')
cumuf = cdf(D)(k)
assert cumuf.dummy_eq(
Sum(Piecewise((1 / n, (ki >= 1) & (ki <= n)), (0, True)), (ki, 1, k)))
assert cumuf.subs({n: 6, k: 2}).doit() == S(1) / 3
t = Dummy('t')
cf = characteristic_function(D)(t)
assert cf.dummy_eq(
Sum(Piecewise((exp(ki * I * t) / n, (ki >= 1) & (ki <= n)), (0, True)),
(ki, 1, n)))
assert cf.subs(
n,
3).doit() == exp(3 * I * t) / 3 + exp(2 * I * t) / 3 + exp(I * t) / 3
mgf = moment_generating_function(D)(t)
assert mgf.dummy_eq(
Sum(Piecewise((exp(ki * t) / n, (ki >= 1) & (ki <= n)), (0, True)),
(ki, 1, n)))
assert mgf.subs(n,
3).doit() == exp(3 * t) / 3 + exp(2 * t) / 3 + exp(t) / 3
