832 lines
30 KiB
Python
832 lines
30 KiB
Python
'''tests for weightstats, compares with replication
|
|
|
|
no failures but needs cleanup
|
|
update 2012-09-09:
|
|
added test after fixing bug in covariance
|
|
TODOs:
|
|
- I do not remember what all the commented out code is doing
|
|
- should be refactored to use generator or inherited tests
|
|
- still gaps in test coverage
|
|
- value/diff in ttest_ind is tested in test_tost.py
|
|
- what about pandas data structures?
|
|
|
|
Author: Josef Perktold
|
|
License: BSD (3-clause)
|
|
|
|
'''
|
|
|
|
import numpy as np
|
|
from scipy import stats
|
|
import pandas as pd
|
|
from numpy.testing import assert_, assert_almost_equal, assert_allclose
|
|
|
|
from statsmodels.stats.weightstats import (DescrStatsW, CompareMeans,
|
|
ttest_ind, ztest, zconfint)
|
|
from statsmodels.tools.testing import Holder
|
|
|
|
|
|
# Mixin for tests against other packages.
|
|
class CheckExternalMixin:
|
|
|
|
@classmethod
|
|
def get_descriptives(cls, ddof=0):
|
|
cls.descriptive = DescrStatsW(cls.data, cls.weights, ddof)
|
|
|
|
# TODO: not a test, belongs elsewhere?
|
|
@classmethod
|
|
def save_data(cls, fname="data.csv"):
|
|
# Utility to get data into another package.
|
|
df = pd.DataFrame(index=np.arange(len(cls.weights)))
|
|
df["weights"] = cls.weights
|
|
if cls.data.ndim == 1:
|
|
df["data1"] = cls.data
|
|
else:
|
|
for k in range(cls.data.shape[1]):
|
|
df["data%d" % (k + 1)] = cls.data[:, k]
|
|
df.to_csv(fname)
|
|
|
|
def test_mean(self):
|
|
mn = self.descriptive.mean
|
|
assert_allclose(mn, self.mean, rtol=1e-4)
|
|
|
|
def test_sum(self):
|
|
sm = self.descriptive.sum
|
|
assert_allclose(sm, self.sum, rtol=1e-4)
|
|
|
|
def test_var(self):
|
|
# Use vardef=wgt option in SAS to match
|
|
var = self.descriptive.var
|
|
assert_allclose(var, self.var, rtol=1e-4)
|
|
|
|
def test_std(self):
|
|
# Use vardef=wgt option in SAS to match
|
|
std = self.descriptive.std
|
|
assert_allclose(std, self.std, rtol=1e-4)
|
|
|
|
def test_sem(self):
|
|
# Use default vardef in SAS to match; only makes sense if
|
|
# weights sum to n.
|
|
if not hasattr(self, "sem"):
|
|
return
|
|
sem = self.descriptive.std_mean
|
|
assert_allclose(sem, self.sem, rtol=1e-4)
|
|
|
|
def test_quantiles(self):
|
|
quant = np.asarray(self.quantiles, dtype=np.float64)
|
|
for return_pandas in False, True:
|
|
qtl = self.descriptive.quantile(self.quantile_probs,
|
|
return_pandas=return_pandas)
|
|
qtl = np.asarray(qtl, dtype=np.float64)
|
|
assert_allclose(qtl, quant, rtol=1e-4)
|
|
|
|
|
|
class TestSim1(CheckExternalMixin):
|
|
# 1d data
|
|
|
|
# Taken from SAS
|
|
mean = 0.401499
|
|
sum = 12.9553441
|
|
var = 1.08022
|
|
std = 1.03933
|
|
quantiles = np.r_[-1.81098, -0.84052, 0.32859, 0.77808, 2.93431]
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(9876789)
|
|
cls.data = np.random.normal(size=20)
|
|
cls.weights = np.random.uniform(0, 3, size=20)
|
|
cls.quantile_probs = np.r_[0, 0.1, 0.5, 0.75, 1]
|
|
cls.get_descriptives()
|
|
|
|
|
|
class TestSim1t(CheckExternalMixin):
|
|
# 1d data with ties
|
|
|
|
# Taken from SAS
|
|
mean = 5.05103296
|
|
sum = 156.573464
|
|
var = 9.9711934
|
|
std = 3.15771965
|
|
quantiles = np.r_[0, 1, 5, 8, 9]
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(9876789)
|
|
cls.data = np.random.randint(0, 10, size=20)
|
|
cls.data[15:20] = cls.data[0:5]
|
|
cls.data[18:20] = cls.data[15:17]
|
|
cls.weights = np.random.uniform(0, 3, size=20)
|
|
cls.quantile_probs = np.r_[0, 0.1, 0.5, 0.75, 1]
|
|
cls.get_descriptives()
|
|
|
|
|
|
class TestSim1n(CheckExternalMixin):
|
|
# 1d data with weights summing to n so we can check the standard
|
|
# error of the mean
|
|
|
|
# Taken from SAS
|
|
mean = -0.3131058
|
|
sum = -6.2621168
|
|
var = 0.49722696
|
|
std = 0.70514322
|
|
sem = 0.15767482
|
|
quantiles = np.r_[-1.61593, -1.45576, -0.24356, 0.16770, 1.18791]
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(4342)
|
|
cls.data = np.random.normal(size=20)
|
|
cls.weights = np.random.uniform(0, 3, size=20)
|
|
cls.weights *= 20 / cls.weights.sum()
|
|
cls.quantile_probs = np.r_[0, 0.1, 0.5, 0.75, 1]
|
|
cls.get_descriptives(1)
|
|
|
|
|
|
class TestSim2(CheckExternalMixin):
|
|
# 2d data
|
|
|
|
# Taken from SAS
|
|
mean = [-0.2170406, -0.2387543]
|
|
sum = [-6.8383999, -7.5225444]
|
|
var = [1.77426344, 0.61933542]
|
|
std = [1.3320148, 0.78697867]
|
|
quantiles = np.column_stack(
|
|
(np.r_[-2.55277, -1.40479, -0.61040, 0.52740, 2.66246],
|
|
np.r_[-1.49263, -1.15403, -0.16231, 0.16464, 1.83062]))
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(2249)
|
|
cls.data = np.random.normal(size=(20, 2))
|
|
cls.weights = np.random.uniform(0, 3, size=20)
|
|
cls.quantile_probs = np.r_[0, 0.1, 0.5, 0.75, 1]
|
|
cls.get_descriptives()
|
|
|
|
|
|
class TestWeightstats:
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(9876789)
|
|
n1, n2 = 20, 20
|
|
m1, m2 = 1, 1.2
|
|
x1 = m1 + np.random.randn(n1)
|
|
x2 = m2 + np.random.randn(n2)
|
|
x1_2d = m1 + np.random.randn(n1, 3)
|
|
x2_2d = m2 + np.random.randn(n2, 3)
|
|
w1 = np.random.randint(1,4, n1)
|
|
w2 = np.random.randint(1,4, n2)
|
|
cls.x1, cls.x2 = x1, x2
|
|
cls.w1, cls.w2 = w1, w2
|
|
cls.x1_2d, cls.x2_2d = x1_2d, x2_2d
|
|
|
|
def test_weightstats_1(self):
|
|
x1, x2 = self.x1, self.x2
|
|
w1, w2 = self.w1, self.w2
|
|
w1_ = 2. * np.ones(len(x1))
|
|
w2_ = 2. * np.ones(len(x2))
|
|
|
|
d1 = DescrStatsW(x1)
|
|
# print ttest_ind(x1, x2)
|
|
# print ttest_ind(x1, x2, usevar='unequal')
|
|
# #print ttest_ind(x1, x2, usevar='unequal')
|
|
# print stats.ttest_ind(x1, x2)
|
|
# print ttest_ind(x1, x2, usevar='unequal', alternative='larger')
|
|
# print ttest_ind(x1, x2, usevar='unequal', alternative='smaller')
|
|
# print ttest_ind(x1, x2, usevar='unequal', weights=(w1_, w2_))
|
|
# print stats.ttest_ind(np.r_[x1, x1], np.r_[x2,x2])
|
|
assert_almost_equal(ttest_ind(x1, x2, weights=(w1_, w2_))[:2],
|
|
stats.ttest_ind(np.r_[x1, x1], np.r_[x2, x2]))
|
|
|
|
def test_weightstats_2(self):
|
|
x1, x2 = self.x1, self.x2
|
|
w1, w2 = self.w1, self.w2
|
|
|
|
d1 = DescrStatsW(x1)
|
|
d1w = DescrStatsW(x1, weights=w1)
|
|
d2w = DescrStatsW(x2, weights=w2)
|
|
x1r = d1w.asrepeats()
|
|
x2r = d2w.asrepeats()
|
|
# print 'random weights'
|
|
# print ttest_ind(x1, x2, weights=(w1, w2))
|
|
# print stats.ttest_ind(x1r, x2r)
|
|
assert_almost_equal(ttest_ind(x1, x2, weights=(w1, w2))[:2],
|
|
stats.ttest_ind(x1r, x2r), 14)
|
|
# not the same as new version with random weights/replication
|
|
# assert x1r.shape[0] == d1w.sum_weights
|
|
# assert x2r.shape[0] == d2w.sum_weights
|
|
|
|
assert_almost_equal(x2r.mean(0), d2w.mean, 14)
|
|
assert_almost_equal(x2r.var(), d2w.var, 14)
|
|
assert_almost_equal(x2r.std(), d2w.std, 14)
|
|
# note: the following is for 1d
|
|
assert_almost_equal(np.cov(x2r, bias=1), d2w.cov, 14)
|
|
# assert_almost_equal(np.corrcoef(np.x2r), d2w.corrcoef, 19)
|
|
# TODO: exception in corrcoef (scalar case)
|
|
|
|
# one-sample tests
|
|
# print d1.ttest_mean(3)
|
|
# print stats.ttest_1samp(x1, 3)
|
|
# print d1w.ttest_mean(3)
|
|
# print stats.ttest_1samp(x1r, 3)
|
|
assert_almost_equal(d1.ttest_mean(3)[:2], stats.ttest_1samp(x1, 3), 11)
|
|
assert_almost_equal(d1w.ttest_mean(3)[:2],
|
|
stats.ttest_1samp(x1r, 3), 11)
|
|
|
|
def test_weightstats_3(self):
|
|
x1_2d, x2_2d = self.x1_2d, self.x2_2d
|
|
w1, w2 = self.w1, self.w2
|
|
|
|
d1w_2d = DescrStatsW(x1_2d, weights=w1)
|
|
d2w_2d = DescrStatsW(x2_2d, weights=w2)
|
|
x1r_2d = d1w_2d.asrepeats()
|
|
x2r_2d = d2w_2d.asrepeats()
|
|
|
|
assert_almost_equal(x2r_2d.mean(0), d2w_2d.mean, 14)
|
|
assert_almost_equal(x2r_2d.var(0), d2w_2d.var, 14)
|
|
assert_almost_equal(x2r_2d.std(0), d2w_2d.std, 14)
|
|
assert_almost_equal(np.cov(x2r_2d.T, bias=1), d2w_2d.cov, 14)
|
|
assert_almost_equal(np.corrcoef(x2r_2d.T), d2w_2d.corrcoef, 14)
|
|
|
|
# print d1w_2d.ttest_mean(3)
|
|
# #scipy.stats.ttest is also vectorized
|
|
# print stats.ttest_1samp(x1r_2d, 3)
|
|
t, p, d = d1w_2d.ttest_mean(3)
|
|
assert_almost_equal([t, p], stats.ttest_1samp(x1r_2d, 3), 11)
|
|
# print [stats.ttest_1samp(xi, 3) for xi in x1r_2d.T]
|
|
cm = CompareMeans(d1w_2d, d2w_2d)
|
|
ressm = cm.ttest_ind()
|
|
resss = stats.ttest_ind(x1r_2d, x2r_2d)
|
|
assert_almost_equal(ressm[:2], resss, 14)
|
|
|
|
# does not work for 2d, levene does not use weights
|
|
# cm = CompareMeans(d1w_2d, d2w_2d)
|
|
# ressm = cm.test_equal_var()
|
|
# resss = stats.levene(x1r_2d, x2r_2d)
|
|
# assert_almost_equal(ressm[:2], resss, 14)
|
|
|
|
def test_weightstats_ddof_tests(self):
|
|
# explicit test that ttest and confint are independent of ddof
|
|
# one sample case
|
|
x1_2d = self.x1_2d
|
|
w1 = self.w1
|
|
|
|
d1w_d0 = DescrStatsW(x1_2d, weights=w1, ddof=0)
|
|
d1w_d1 = DescrStatsW(x1_2d, weights=w1, ddof=1)
|
|
d1w_d2 = DescrStatsW(x1_2d, weights=w1, ddof=2)
|
|
|
|
# check confint independent of user ddof
|
|
res0 = d1w_d0.ttest_mean()
|
|
res1 = d1w_d1.ttest_mean()
|
|
res2 = d1w_d2.ttest_mean()
|
|
# concatenate into one array with np.r_
|
|
assert_almost_equal(np.r_[res1], np.r_[res0], 14)
|
|
assert_almost_equal(np.r_[res2], np.r_[res0], 14)
|
|
|
|
res0 = d1w_d0.ttest_mean(0.5)
|
|
res1 = d1w_d1.ttest_mean(0.5)
|
|
res2 = d1w_d2.ttest_mean(0.5)
|
|
assert_almost_equal(np.r_[res1], np.r_[res0], 14)
|
|
assert_almost_equal(np.r_[res2], np.r_[res0], 14)
|
|
|
|
# check confint independent of user ddof
|
|
res0 = d1w_d0.tconfint_mean()
|
|
res1 = d1w_d1.tconfint_mean()
|
|
res2 = d1w_d2.tconfint_mean()
|
|
assert_almost_equal(res1, res0, 14)
|
|
assert_almost_equal(res2, res0, 14)
|
|
|
|
def test_comparemeans_convenient_interface(self):
|
|
x1_2d, x2_2d = self.x1_2d, self.x2_2d
|
|
d1 = DescrStatsW(x1_2d)
|
|
d2 = DescrStatsW(x2_2d)
|
|
cm1 = CompareMeans(d1, d2)
|
|
|
|
# smoke test for summary
|
|
from statsmodels.iolib.table import SimpleTable
|
|
for use_t in [True, False]:
|
|
for usevar in ['pooled', 'unequal']:
|
|
smry = cm1.summary(use_t=use_t, usevar=usevar)
|
|
assert_(isinstance(smry, SimpleTable))
|
|
|
|
# test for from_data method
|
|
cm2 = CompareMeans.from_data(x1_2d, x2_2d)
|
|
assert_(str(cm1.summary()) == str(cm2.summary()))
|
|
|
|
def test_comparemeans_convenient_interface_1d(self):
|
|
# same as above for 2d, just use 1d data instead
|
|
x1_2d, x2_2d = self.x1, self.x2
|
|
d1 = DescrStatsW(x1_2d)
|
|
d2 = DescrStatsW(x2_2d)
|
|
cm1 = CompareMeans(d1, d2)
|
|
|
|
# smoke test for summary
|
|
from statsmodels.iolib.table import SimpleTable
|
|
for use_t in [True, False]:
|
|
for usevar in ['pooled', 'unequal']:
|
|
smry = cm1.summary(use_t=use_t, usevar=usevar)
|
|
assert_(isinstance(smry, SimpleTable))
|
|
|
|
# test for from_data method
|
|
cm2 = CompareMeans.from_data(x1_2d, x2_2d)
|
|
assert_(str(cm1.summary()) == str(cm2.summary()))
|
|
|
|
|
|
class CheckWeightstats1dMixin:
|
|
|
|
def test_basic(self):
|
|
x1r = self.x1r
|
|
d1w = self.d1w
|
|
|
|
assert_almost_equal(x1r.mean(0), d1w.mean, 14)
|
|
assert_almost_equal(x1r.var(0, ddof=d1w.ddof), d1w.var, 14)
|
|
assert_almost_equal(x1r.std(0, ddof=d1w.ddof), d1w.std, 14)
|
|
var1 = d1w.var_ddof(ddof=1)
|
|
assert_almost_equal(x1r.var(0, ddof=1), var1, 14)
|
|
std1 = d1w.std_ddof(ddof=1)
|
|
assert_almost_equal(x1r.std(0, ddof=1), std1, 14)
|
|
|
|
assert_almost_equal(np.cov(x1r.T, bias=1-d1w.ddof), d1w.cov, 14)
|
|
|
|
# assert_almost_equal(np.corrcoef(x1r.T), d1w.corrcoef, 14)
|
|
|
|
def test_ttest(self):
|
|
x1r = self.x1r
|
|
d1w = self.d1w
|
|
assert_almost_equal(d1w.ttest_mean(3)[:2],
|
|
stats.ttest_1samp(x1r, 3), 11)
|
|
|
|
# def
|
|
# assert_almost_equal(ttest_ind(x1, x2, weights=(w1, w2))[:2],
|
|
# stats.ttest_ind(x1r, x2r), 14)
|
|
|
|
def test_ttest_2sample(self):
|
|
x1, x2 = self.x1, self.x2
|
|
x1r, x2r = self.x1r, self.x2r
|
|
w1, w2 = self.w1, self.w2
|
|
|
|
# Note: stats.ttest_ind handles 2d/nd arguments
|
|
res_sp = stats.ttest_ind(x1r, x2r)
|
|
assert_almost_equal(ttest_ind(x1, x2, weights=(w1, w2))[:2],
|
|
res_sp, 14)
|
|
|
|
# check correct ttest independent of user ddof
|
|
cm = CompareMeans(DescrStatsW(x1, weights=w1, ddof=0),
|
|
DescrStatsW(x2, weights=w2, ddof=1))
|
|
assert_almost_equal(cm.ttest_ind()[:2], res_sp, 14)
|
|
|
|
cm = CompareMeans(DescrStatsW(x1, weights=w1, ddof=1),
|
|
DescrStatsW(x2, weights=w2, ddof=2))
|
|
assert_almost_equal(cm.ttest_ind()[:2], res_sp, 14)
|
|
|
|
cm0 = CompareMeans(DescrStatsW(x1, weights=w1, ddof=0),
|
|
DescrStatsW(x2, weights=w2, ddof=0))
|
|
cm1 = CompareMeans(DescrStatsW(x1, weights=w1, ddof=0),
|
|
DescrStatsW(x2, weights=w2, ddof=1))
|
|
cm2 = CompareMeans(DescrStatsW(x1, weights=w1, ddof=1),
|
|
DescrStatsW(x2, weights=w2, ddof=2))
|
|
|
|
res0 = cm0.ttest_ind(usevar='unequal')
|
|
res1 = cm1.ttest_ind(usevar='unequal')
|
|
res2 = cm2.ttest_ind(usevar='unequal')
|
|
assert_almost_equal(res1, res0, 14)
|
|
assert_almost_equal(res2, res0, 14)
|
|
|
|
# check confint independent of user ddof
|
|
res0 = cm0.tconfint_diff(usevar='pooled')
|
|
res1 = cm1.tconfint_diff(usevar='pooled')
|
|
res2 = cm2.tconfint_diff(usevar='pooled')
|
|
assert_almost_equal(res1, res0, 14)
|
|
assert_almost_equal(res2, res0, 14)
|
|
|
|
res0 = cm0.tconfint_diff(usevar='unequal')
|
|
res1 = cm1.tconfint_diff(usevar='unequal')
|
|
res2 = cm2.tconfint_diff(usevar='unequal')
|
|
assert_almost_equal(res1, res0, 14)
|
|
assert_almost_equal(res2, res0, 14)
|
|
|
|
def test_confint_mean(self):
|
|
# compare confint_mean with ttest
|
|
d1w = self.d1w
|
|
alpha = 0.05
|
|
low, upp = d1w.tconfint_mean()
|
|
t, p, d = d1w.ttest_mean(low)
|
|
assert_almost_equal(p, alpha * np.ones(p.shape), 8)
|
|
t, p, d = d1w.ttest_mean(upp)
|
|
assert_almost_equal(p, alpha * np.ones(p.shape), 8)
|
|
t, p, d = d1w.ttest_mean(np.vstack((low, upp)))
|
|
assert_almost_equal(p, alpha * np.ones(p.shape), 8)
|
|
|
|
|
|
class CheckWeightstats2dMixin(CheckWeightstats1dMixin):
|
|
|
|
def test_corr(self):
|
|
x1r = self.x1r
|
|
d1w = self.d1w
|
|
|
|
assert_almost_equal(np.corrcoef(x1r.T), d1w.corrcoef, 14)
|
|
|
|
|
|
class TestWeightstats1d_ddof(CheckWeightstats1dMixin):
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(9876789)
|
|
n1, n2 = 20, 20
|
|
m1, m2 = 1, 1.2
|
|
x1 = m1 + np.random.randn(n1, 1)
|
|
x2 = m2 + np.random.randn(n2, 1)
|
|
w1 = np.random.randint(1, 4, n1)
|
|
w2 = np.random.randint(1, 4, n2)
|
|
|
|
cls.x1, cls.x2 = x1, x2
|
|
cls.w1, cls.w2 = w1, w2
|
|
cls.d1w = DescrStatsW(x1, weights=w1, ddof=1)
|
|
cls.d2w = DescrStatsW(x2, weights=w2, ddof=1)
|
|
cls.x1r = cls.d1w.asrepeats()
|
|
cls.x2r = cls.d2w.asrepeats()
|
|
|
|
|
|
class TestWeightstats2d(CheckWeightstats2dMixin):
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(9876789)
|
|
n1, n2 = 20, 20
|
|
m1, m2 = 1, 1.2
|
|
x1 = m1 + np.random.randn(n1, 3)
|
|
x2 = m2 + np.random.randn(n2, 3)
|
|
w1 = np.random.randint(1, 4, n1)
|
|
w2 = np.random.randint(1, 4, n2)
|
|
cls.x1, cls.x2 = x1, x2
|
|
cls.w1, cls.w2 = w1, w2
|
|
|
|
cls.d1w = DescrStatsW(x1, weights=w1)
|
|
cls.d2w = DescrStatsW(x2, weights=w2)
|
|
cls.x1r = cls.d1w.asrepeats()
|
|
cls.x2r = cls.d2w.asrepeats()
|
|
|
|
|
|
class TestWeightstats2d_ddof(CheckWeightstats2dMixin):
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(9876789)
|
|
n1, n2 = 20, 20
|
|
m1, m2 = 1, 1.2
|
|
x1 = m1 + np.random.randn(n1, 3)
|
|
x2 = m2 + np.random.randn(n2, 3)
|
|
w1 = np.random.randint(1, 4, n1)
|
|
w2 = np.random.randint(1, 4, n2)
|
|
|
|
cls.x1, cls.x2 = x1, x2
|
|
cls.w1, cls.w2 = w1, w2
|
|
cls.d1w = DescrStatsW(x1, weights=w1, ddof=1)
|
|
cls.d2w = DescrStatsW(x2, weights=w2, ddof=1)
|
|
cls.x1r = cls.d1w.asrepeats()
|
|
cls.x2r = cls.d2w.asrepeats()
|
|
|
|
|
|
class TestWeightstats2d_nobs(CheckWeightstats2dMixin):
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
np.random.seed(9876789)
|
|
n1, n2 = 20, 30
|
|
m1, m2 = 1, 1.2
|
|
x1 = m1 + np.random.randn(n1, 3)
|
|
x2 = m2 + np.random.randn(n2, 3)
|
|
w1 = np.random.randint(1, 4, n1)
|
|
w2 = np.random.randint(1, 4, n2)
|
|
|
|
cls.x1, cls.x2 = x1, x2
|
|
cls.w1, cls.w2 = w1, w2
|
|
cls.d1w = DescrStatsW(x1, weights=w1, ddof=0)
|
|
cls.d2w = DescrStatsW(x2, weights=w2, ddof=1)
|
|
cls.x1r = cls.d1w.asrepeats()
|
|
cls.x2r = cls.d2w.asrepeats()
|
|
|
|
|
|
def test_ttest_ind_with_uneq_var():
|
|
|
|
# from scipy
|
|
# check vs. R
|
|
a = (1, 2, 3)
|
|
b = (1.1, 2.9, 4.2)
|
|
pr = 0.53619490753126731
|
|
tr = -0.68649512735572582
|
|
t, p, df = ttest_ind(a, b, usevar='unequal')
|
|
assert_almost_equal([t, p], [tr, pr], 13)
|
|
|
|
a = (1, 2, 3, 4)
|
|
pr = 0.84354139131608286
|
|
tr = -0.2108663315950719
|
|
t, p, df = ttest_ind(a, b, usevar='unequal')
|
|
assert_almost_equal([t, p], [tr, pr], 13)
|
|
|
|
|
|
def test_ztest_ztost():
|
|
# compare weightstats with separately tested proportion ztest ztost
|
|
import statsmodels.stats.proportion as smprop
|
|
|
|
x1 = [0, 1]
|
|
w1 = [5, 15]
|
|
|
|
res2 = smprop.proportions_ztest(15, 20., value=0.5)
|
|
d1 = DescrStatsW(x1, w1)
|
|
res1 = d1.ztest_mean(0.5)
|
|
assert_allclose(res1, res2, rtol=0.03, atol=0.003)
|
|
|
|
d2 = DescrStatsW(x1, np.array(w1)*21./20)
|
|
res1 = d2.ztest_mean(0.5)
|
|
assert_almost_equal(res1, res2, decimal=12)
|
|
|
|
res1 = d2.ztost_mean(0.4, 0.6)
|
|
res2 = smprop.proportions_ztost(15, 20., 0.4, 0.6)
|
|
assert_almost_equal(res1[0], res2[0], decimal=12)
|
|
|
|
x2 = [0, 1]
|
|
w2 = [10, 10]
|
|
# d2 = DescrStatsW(x1, np.array(w1)*21./20)
|
|
d2 = DescrStatsW(x2, w2)
|
|
res1 = ztest(d1.asrepeats(), d2.asrepeats())
|
|
res2 = smprop.proportions_chisquare(np.asarray([15, 10]),
|
|
np.asarray([20., 20]))
|
|
# TODO: check this is this difference expected?, see test_proportion
|
|
assert_allclose(res1[1], res2[1], rtol=0.03)
|
|
|
|
res1a = CompareMeans(d1, d2).ztest_ind()
|
|
assert_allclose(res1a[1], res2[1], rtol=0.03)
|
|
assert_almost_equal(res1a, res1, decimal=12)
|
|
|
|
|
|
# test for ztest and z confidence interval against R BSDA z.test
|
|
# Note: I needed to calculate the pooled standard deviation for R
|
|
# std = np.std(np.concatenate((x-x.mean(),y-y.mean())), ddof=2)
|
|
|
|
# > zt = z.test(x, sigma.x=0.57676142668828667, y, sigma.y=0.57676142668828667)
|
|
# > cat_items(zt, "ztest.")
|
|
ztest_ = Holder()
|
|
ztest_.statistic = 6.55109865675183
|
|
ztest_.p_value = 5.711530850508982e-11
|
|
ztest_.conf_int = np.array([1.230415246535603, 2.280948389828034])
|
|
ztest_.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_.null_value = 0
|
|
ztest_.alternative = 'two.sided'
|
|
ztest_.method = 'Two-sample z-Test'
|
|
ztest_.data_name = 'x and y'
|
|
# > zt = z.test(x, sigma.x=0.57676142668828667, y,
|
|
# sigma.y=0.57676142668828667, alternative="less")
|
|
# > cat_items(zt, "ztest_smaller.")
|
|
ztest_smaller = Holder()
|
|
ztest_smaller.statistic = 6.55109865675183
|
|
ztest_smaller.p_value = 0.999999999971442
|
|
ztest_smaller.conf_int = np.array([np.nan, 2.196499421109045])
|
|
ztest_smaller.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_smaller.null_value = 0
|
|
ztest_smaller.alternative = 'less'
|
|
ztest_smaller.method = 'Two-sample z-Test'
|
|
ztest_smaller.data_name = 'x and y'
|
|
# > zt = z.test(x, sigma.x=0.57676142668828667, y,
|
|
# sigma.y=0.57676142668828667, alternative="greater")
|
|
# > cat_items(zt, "ztest_larger.")
|
|
ztest_larger = Holder()
|
|
ztest_larger.statistic = 6.55109865675183
|
|
ztest_larger.p_value = 2.855760072861813e-11
|
|
ztest_larger.conf_int = np.array([1.314864215254592, np.nan])
|
|
ztest_larger.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_larger.null_value = 0
|
|
ztest_larger.alternative = 'greater'
|
|
ztest_larger.method = 'Two-sample z-Test'
|
|
ztest_larger.data_name = 'x and y'
|
|
|
|
|
|
# > zt = z.test(x, sigma.x=0.57676142668828667, y,
|
|
# sigma.y=0.57676142668828667, mu=1, alternative="two.sided")
|
|
# > cat_items(zt, "ztest_mu.")
|
|
ztest_mu = Holder()
|
|
ztest_mu.statistic = 2.81972854805176
|
|
ztest_mu.p_value = 0.00480642898427981
|
|
ztest_mu.conf_int = np.array([1.230415246535603, 2.280948389828034])
|
|
ztest_mu.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_mu.null_value = 1
|
|
ztest_mu.alternative = 'two.sided'
|
|
ztest_mu.method = 'Two-sample z-Test'
|
|
ztest_mu.data_name = 'x and y'
|
|
|
|
# > zt = z.test(x, sigma.x=0.57676142668828667, y,
|
|
# sigma.y=0.57676142668828667, mu=1, alternative="greater")
|
|
# > cat_items(zt, "ztest_larger_mu.")
|
|
ztest_larger_mu = Holder()
|
|
ztest_larger_mu.statistic = 2.81972854805176
|
|
ztest_larger_mu.p_value = 0.002403214492139871
|
|
ztest_larger_mu.conf_int = np.array([1.314864215254592, np.nan])
|
|
ztest_larger_mu.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_larger_mu.null_value = 1
|
|
ztest_larger_mu.alternative = 'greater'
|
|
ztest_larger_mu.method = 'Two-sample z-Test'
|
|
ztest_larger_mu.data_name = 'x and y'
|
|
|
|
# > zt = z.test(x, sigma.x=0.57676142668828667, y,
|
|
# sigma.y=0.57676142668828667, mu=2, alternative="less")
|
|
# > cat_items(zt, "ztest_smaller_mu.")
|
|
ztest_smaller_mu = Holder()
|
|
ztest_smaller_mu.statistic = -0.911641560648313
|
|
ztest_smaller_mu.p_value = 0.1809787183191324
|
|
ztest_smaller_mu.conf_int = np.array([np.nan, 2.196499421109045])
|
|
ztest_smaller_mu.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_smaller_mu.null_value = 2
|
|
ztest_smaller_mu.alternative = 'less'
|
|
ztest_smaller_mu.method = 'Two-sample z-Test'
|
|
ztest_smaller_mu.data_name = 'x and y'
|
|
|
|
# > zt = z.test(x, sigma.x=0.46436662631627995, mu=6.4,
|
|
# alternative="two.sided")
|
|
# > cat_items(zt, "ztest_mu_1s.")
|
|
ztest_mu_1s = Holder()
|
|
ztest_mu_1s.statistic = 4.415212090914452
|
|
ztest_mu_1s.p_value = 1.009110038015147e-05
|
|
ztest_mu_1s.conf_int = np.array([6.74376372125119, 7.29259991511245])
|
|
ztest_mu_1s.estimate = 7.01818181818182
|
|
ztest_mu_1s.null_value = 6.4
|
|
ztest_mu_1s.alternative = 'two.sided'
|
|
ztest_mu_1s.method = 'One-sample z-Test'
|
|
ztest_mu_1s.data_name = 'x'
|
|
|
|
# > zt = z.test(x, sigma.x=0.46436662631627995, mu=7.4, alternative="less")
|
|
# > cat_items(zt, "ztest_smaller_mu_1s.")
|
|
ztest_smaller_mu_1s = Holder()
|
|
ztest_smaller_mu_1s.statistic = -2.727042762035397
|
|
ztest_smaller_mu_1s.p_value = 0.00319523783881176
|
|
ztest_smaller_mu_1s.conf_int = np.array([np.nan, 7.248480744895716])
|
|
ztest_smaller_mu_1s.estimate = 7.01818181818182
|
|
ztest_smaller_mu_1s.null_value = 7.4
|
|
ztest_smaller_mu_1s.alternative = 'less'
|
|
ztest_smaller_mu_1s.method = 'One-sample z-Test'
|
|
ztest_smaller_mu_1s.data_name = 'x'
|
|
|
|
# > zt = z.test(x, sigma.x=0.46436662631627995, mu=6.4, alternative="greater")
|
|
# > cat_items(zt, "ztest_greater_mu_1s.")
|
|
ztest_larger_mu_1s = Holder()
|
|
ztest_larger_mu_1s.statistic = 4.415212090914452
|
|
ztest_larger_mu_1s.p_value = 5.045550190097003e-06
|
|
ztest_larger_mu_1s.conf_int = np.array([6.78788289146792, np.nan])
|
|
ztest_larger_mu_1s.estimate = 7.01818181818182
|
|
ztest_larger_mu_1s.null_value = 6.4
|
|
ztest_larger_mu_1s.alternative = 'greater'
|
|
ztest_larger_mu_1s.method = 'One-sample z-Test'
|
|
ztest_larger_mu_1s.data_name = 'x'
|
|
|
|
# > zt = z.test(x, sigma.x=0.46436662631627995, y, sigma.y=0.7069805008424409)
|
|
# > cat_items(zt, "ztest_unequal.")
|
|
ztest_unequal = Holder()
|
|
ztest_unequal.statistic = 6.12808151466544
|
|
ztest_unequal.p_value = 8.89450168270109e-10
|
|
ztest_unequal.conf_int = np.array([1.19415646579981, 2.31720717056382])
|
|
ztest_unequal.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_unequal.null_value = 0
|
|
ztest_unequal.alternative = 'two.sided'
|
|
ztest_unequal.usevar = 'unequal'
|
|
ztest_unequal.method = 'Two-sample z-Test'
|
|
ztest_unequal.data_name = 'x and y'
|
|
|
|
# > zt = z.test(x, sigma.x=0.46436662631627995, y, sigma.y=0.7069805008424409, alternative="less")
|
|
# > cat_items(zt, "ztest_smaller_unequal.")
|
|
ztest_smaller_unequal = Holder()
|
|
ztest_smaller_unequal.statistic = 6.12808151466544
|
|
ztest_smaller_unequal.p_value = 0.999999999555275
|
|
ztest_smaller_unequal.conf_int = np.array([np.nan, 2.22692874913371])
|
|
ztest_smaller_unequal.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_smaller_unequal.null_value = 0
|
|
ztest_smaller_unequal.alternative = 'less'
|
|
ztest_smaller_unequal.usevar = 'unequal'
|
|
ztest_smaller_unequal.method = 'Two-sample z-Test'
|
|
ztest_smaller_unequal.data_name = 'x and y'
|
|
|
|
# > zt = z.test(x, sigma.x=0.46436662631627995, y, sigma.y=0.7069805008424409, alternative="greater")
|
|
# > cat_items(zt, "ztest_larger_unequal.")
|
|
ztest_larger_unequal = Holder()
|
|
ztest_larger_unequal.statistic = 6.12808151466544
|
|
ztest_larger_unequal.p_value = 4.44725034576265e-10
|
|
ztest_larger_unequal.conf_int = np.array([1.28443488722992, np.nan])
|
|
ztest_larger_unequal.estimate = np.array([7.01818181818182, 5.2625])
|
|
ztest_larger_unequal.null_value = 0
|
|
ztest_larger_unequal.alternative = 'greater'
|
|
ztest_larger_unequal.usevar = 'unequal'
|
|
ztest_larger_unequal.method = 'Two-sample z-Test'
|
|
ztest_larger_unequal.data_name = 'x and y'
|
|
|
|
|
|
alternatives = {'less': 'smaller',
|
|
'greater': 'larger',
|
|
'two.sided': 'two-sided'}
|
|
|
|
|
|
class TestZTest:
|
|
# all examples use the same data
|
|
# no weights used in tests
|
|
|
|
@classmethod
|
|
def setup_class(cls):
|
|
cls.x1 = np.array([7.8, 6.6, 6.5, 7.4, 7.3, 7., 6.4,
|
|
7.1, 6.7, 7.6, 6.8])
|
|
cls.x2 = np.array([4.5, 5.4, 6.1, 6.1, 5.4, 5., 4.1, 5.5])
|
|
cls.d1 = DescrStatsW(cls.x1)
|
|
cls.d2 = DescrStatsW(cls.x2)
|
|
cls.cm = CompareMeans(cls.d1, cls.d2)
|
|
|
|
def test(self):
|
|
x1, x2 = self.x1, self.x2
|
|
cm = self.cm
|
|
|
|
# tc : test cases
|
|
for tc in [ztest_, ztest_smaller, ztest_larger,
|
|
ztest_mu, ztest_smaller_mu, ztest_larger_mu]:
|
|
|
|
zstat, pval = ztest(x1, x2, value=tc.null_value,
|
|
alternative=alternatives[tc.alternative])
|
|
assert_allclose(zstat, tc.statistic, rtol=1e-10)
|
|
assert_allclose(pval, tc.p_value, rtol=1e-10, atol=1e-16)
|
|
|
|
zstat, pval = cm.ztest_ind(value=tc.null_value,
|
|
alternative=alternatives[tc.alternative])
|
|
assert_allclose(zstat, tc.statistic, rtol=1e-10)
|
|
assert_allclose(pval, tc.p_value, rtol=1e-10, atol=1e-16)
|
|
|
|
# overwrite nan in R's confint
|
|
tc_conf_int = tc.conf_int.copy()
|
|
if np.isnan(tc_conf_int[0]):
|
|
tc_conf_int[0] = - np.inf
|
|
if np.isnan(tc_conf_int[1]):
|
|
tc_conf_int[1] = np.inf
|
|
|
|
# Note: value is shifting our confidence interval in zconfint
|
|
ci = zconfint(x1, x2, value=0,
|
|
alternative=alternatives[tc.alternative])
|
|
assert_allclose(ci, tc_conf_int, rtol=1e-10)
|
|
|
|
ci = cm.zconfint_diff(alternative=alternatives[tc.alternative])
|
|
assert_allclose(ci, tc_conf_int, rtol=1e-10)
|
|
|
|
ci = zconfint(x1, x2, value=tc.null_value,
|
|
alternative=alternatives[tc.alternative])
|
|
assert_allclose(ci, tc_conf_int - tc.null_value, rtol=1e-10)
|
|
|
|
# unequal variances
|
|
for tc in [ztest_unequal, ztest_smaller_unequal, ztest_larger_unequal]:
|
|
zstat, pval = ztest(x1, x2, value=tc.null_value,
|
|
alternative=alternatives[tc.alternative],
|
|
usevar="unequal")
|
|
assert_allclose(zstat, tc.statistic, rtol=1e-10)
|
|
assert_allclose(pval, tc.p_value, rtol=1e-10, atol=1e-16)
|
|
|
|
# 1 sample test copy-paste
|
|
d1 = self.d1
|
|
for tc in [ztest_mu_1s, ztest_smaller_mu_1s, ztest_larger_mu_1s]:
|
|
zstat, pval = ztest(x1, value=tc.null_value,
|
|
alternative=alternatives[tc.alternative])
|
|
assert_allclose(zstat, tc.statistic, rtol=1e-10)
|
|
assert_allclose(pval, tc.p_value, rtol=1e-10, atol=1e-16)
|
|
|
|
zstat, pval = d1.ztest_mean(value=tc.null_value,
|
|
alternative=alternatives[tc.alternative])
|
|
assert_allclose(zstat, tc.statistic, rtol=1e-10)
|
|
assert_allclose(pval, tc.p_value, rtol=1e-10, atol=1e-16)
|
|
|
|
# overwrite nan in R's confint
|
|
tc_conf_int = tc.conf_int.copy()
|
|
if np.isnan(tc_conf_int[0]):
|
|
tc_conf_int[0] = - np.inf
|
|
if np.isnan(tc_conf_int[1]):
|
|
tc_conf_int[1] = np.inf
|
|
|
|
# Note: value is shifting our confidence interval in zconfint
|
|
ci = zconfint(x1, value=0,
|
|
alternative=alternatives[tc.alternative])
|
|
assert_allclose(ci, tc_conf_int, rtol=1e-10)
|
|
|
|
ci = d1.zconfint_mean(alternative=alternatives[tc.alternative])
|
|
assert_allclose(ci, tc_conf_int, rtol=1e-10)
|
|
|
|
|
|
def test_weightstats_len_1():
|
|
x1 = [1]
|
|
w1 = [1]
|
|
d1 = DescrStatsW(x1, w1)
|
|
assert (d1.quantile([0.0, 0.5, 1.0]) == 1).all()
|
|
|
|
|
|
def test_weightstats_2d_w1():
|
|
x1 = [[1], [2]]
|
|
w1 = [[1], [2]]
|
|
d1 = DescrStatsW(x1, w1)
|
|
print(len(np.array(w1).shape))
|
|
assert (d1.quantile([0.5, 1.0]) == 2).all().all()
|
|
|
|
|
|
def test_weightstats_2d_w2():
|
|
x1 = [[1]]
|
|
w1 = [[1]]
|
|
d1 = DescrStatsW(x1, w1)
|
|
assert (d1.quantile([0, 0.5, 1.0]) == 1).all().all()
|