def test_likelihood(n_rows, n_cols, n_cats, cat_sep, n_models, n_iter):
dg, engine = gen_data_and_engine(n_rows, n_cols, n_cats, cat_sep, n_models,
n_iter)
cols = list(range(n_cols))
ttl_base = "CSD_LK-r%d-c%d-k%d-s%1.2f-m%d-i%d_" \
% (n_rows, n_cols, n_cats, cat_sep, n_models, n_iter,)
for col in cols:
x = np.arange(5, dtype=int)
l_dg = dg.log_likelihood(x, col)
l_eng = engine.probability(x[:, np.newaxis], [col])
r, _ = pearsonr(l_dg, l_eng)
inftest_bar(np.exp(l_dg), np.exp(l_eng), ttl_base + 'COL-%d' % col,
RESDIR)
assert abs(r - 1) < .1
def test_ks(n_rows, n_cols, n_cats, cat_sep, n_models, n_iter):
dg, engine = gen_data_and_engine(n_rows, n_cols, n_cats, cat_sep, n_models,
n_iter)
cols = list(range(n_cols))
ttl_base = "CSD_X2-r%d-c%d-k%d-s%1.2f-m%d-i%d_" \
% (n_rows, n_cols, n_cats, cat_sep, n_models, n_iter,)
if n_cols == 1:
y = sum(p['alpha'] for p in dg.params[0]) / n_cats
x = np.array(engine.sample([0], n=n_rows).flatten(), dtype=int)
x = np.bincount(x, minlength=5) / n_rows
d, p = chisquare(y, x)
inftest_bar(y, x, ttl_base + 'COL-0', RESDIR)
assert p > .05
else:
for col in cols:
y = sum(p['alpha'] for p in dg.params[0]) / n_cats
x = np.array(engine.sample([0], n=n_rows).flatten(), dtype=int)
x = np.bincount(x, minlength=5) / n_rows
d, p = chisquare(y, x)
inftest_bar(y, x, ttl_base + 'COL-%d' % col, RESDIR)
assert p > .05
