def test_bootstrap_simple_mean(self):
data = np.ones(10)
results = statistics.bootstrap(
data,
np.mean,
bootstrap_iterations=5,
)
self.assertListEqual(results, [1 for _ in range(5)])
def test_bootstrap_sample_size_fraction(self):
data = np.ones(10)
results = statistics.bootstrap(
data,
len,
bootstrap_sample_size=0.5,
bootstrap_iterations=2,
)
for i in range(2):
self.assertTrue(results[i] == 5, f"{i} has length {results[i]}")
def test_bootstrap_min_sample_size_integer(self):
data = np.ones(100)
results = statistics.bootstrap(
data,
len,
bootstrap_min_sample_size=3,
bootstrap_iterations=10,
sample_method="integer",
)
for i in range(10):
self.assertTrue(results[i] >= 3, f"{i} has length {results[i]}")
def test_bootstrap_null(self):
def func(_, a, b=1):
return a + b
data = np.ones(10)
results = statistics.bootstrap(
data,
func,
func_args=(1,),
func_kws={"b": 1},
bootstrap_iterations=5,
)
self.assertListEqual(results, [2 for _ in range(5)])
def test_bootstrap_custom_sampler(self):
data = np.arange(10)
def sample_method(_, number):
return np.arange(0, number)
results = statistics.bootstrap(
data,
np.mean,
bootstrap_sample_size=5, # select first 5 every time
bootstrap_iterations=4, # number of iteractions
sample_method=sample_method,
)
# first three are [0, 1, 2, 3, 4] so mean is 2
self.assertListEqual(results, [2 for _ in range(4)])