def frequencies(Y, X, sub=None, title="{Yname} Frequencies"):
"""
Display frequency of occurrence of all categories in Y in the cells
defined by X.
Y: vactor whose ferquencies are of interest
X: model defining cells for which frequencies are displayed
"""
Y = _data.asfactor(Y)
X = _data.asfactor(X)
cells = _structure.celltable(Y, X, sub=sub)
Y_categories = cells.Y.cells.keys()
# header
n_Y_categories = len(Y_categories)
table = textab.Table('l' * (n_Y_categories+1))
# header line 1
table.cell()
table.cell(Y.name, width=n_Y_categories, just='c')
table.midrule(span=(2, n_Y_categories+1))
# header line 2
table.cell(X.name)
for cat_id in Y_categories:
table.cell(cells.Y.cells[cat_id])
table.midrule()
# body
for cell_id in cells.indexes:
table.cell(cells.cells[cell_id])
data = cells.data[cell_id]
for cat_id in Y_categories:
n = np.sum(cat_id == data)
table.cell(n)
# title
if title:
title = title.format(Yname=Y.name.capitalize())
table.title(title)
return table
def __init__(self, Y, X, match=None, sub=None,
samples=1000, replacement=True,
title="Bootstrapped Pairwise Tests"):
Y = asvar(Y, sub)
X = asfactor(X, sub)
assert len(Y) == len(X), "dataset length mismatch"
if match:
if sub is not None:
match = match[sub]
assert len(match) == len(Y), "dataset length mismatch"
# prepare data container
resampled = np.empty((samples + 1, len(Y))) # sample X subject within category
resampled[0] = Y.x
# fill resampled
for i, Y_resampled in _resample(Y, unit=match, samples=samples,
replacement=replacement):
resampled[i + 1] = Y_resampled.x
self.resampled = resampled
cells = X.cells
n_groups = len(cells)
if match:
# if there are several values per X%match cell, take the average
# T: indexes to transform Y.x to [X%match, value]-array
match_cell_ids = match.cells
group_size = len(match_cell_ids)
T = None; i = 0
for X_cell in cells:
for match_cell in match_cell_ids:
source_indexes = np.where((X == X_cell) * (match == match_cell))[0]
if T is None:
n_cells = n_groups * group_size
T = np.empty((n_cells, len(source_indexes)), dtype=int)
T[i, :] = source_indexes
i += 1
if T.shape[1] == 1:
T = T[:, 0]
ordered = resampled[:, T]
else:
ordered = resampled[:, T].mean(axis=2)
self.ordered = ordered
# t-tests
n_comparisons = sum(range(n_groups))
t = np.empty((samples + 1, n_comparisons))
comp_names = []
one_group = np.arange(group_size)
groups = [one_group + i * group_size for i in range(n_groups)]
for i, (g1, g2) in enumerate(itertools.combinations(range(n_groups), 2)):
group_1 = groups[g1]
group_2 = groups[g2]
diffs = ordered[:, group_1] - ordered[:, group_2]
t[:, i] = np.mean(diffs, axis=1) * np.sqrt(group_size) / np.std(diffs, axis=1, ddof=1)
comp_names.append(' - '.join((cells[g1], cells[g2])))
self.diffs = diffs
self.t_resampled = np.max(np.abs(t[1:]), axis=1)
self.t = t = t[0]
else:
raise NotImplementedError
self._Y = Y
self._X = X
self._group_names = cells
self._group_data = np.array([ordered[0, g] for g in groups])
self._group_size = group_size
self._df = group_size - 1
self._match = match
self._n_samples = samples
self._replacement = replacement
self._comp_names = comp_names
self._p_parametric = self.test_param(t)
self._p_boot = self.test_boot(t)
self.title = title