-
Notifications
You must be signed in to change notification settings - Fork 45
/
_classes.py
1414 lines (1106 loc) · 48.5 KB
/
_classes.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#!/usr/bin/python
# -*-coding: utf-8 -*-
# Author: Joses Ho
# Email : joseshowh@gmail.com
class Dabest(object):
"""
Class for estimation statistics and plots.
"""
def __init__(self, data, idx, x, y, paired, id_col, ci, resamples,
random_seed):
"""
Parses and stores pandas DataFrames in preparation for estimation
statistics.
"""
# Import standard data science libraries.
import numpy as np
import pandas as pd
import seaborn as sns
self.__ci = ci
self.__data = data
self.__idx = idx
self.__id_col = id_col
self.__is_paired = paired
self.__resamples = resamples
self.__random_seed = random_seed
# Make a copy of the data, so we don't make alterations to it.
data_in = data.copy()
# data_in.reset_index(inplace=True)
# data_in_index_name = data_in.index.name
# Determine the kind of estimation plot we need to produce.
if all([isinstance(i, str) for i in idx]):
# flatten out idx.
all_plot_groups = pd.unique([t for t in idx]).tolist()
if len(idx) > len(all_plot_groups):
err0 = '`idx` contains duplicated groups. Please remove any duplicates and try again.'
raise ValueError(err0)
# We need to re-wrap this idx inside another tuple so as to
# easily loop thru each pairwise group later on.
self.__idx = (idx,)
elif all([isinstance(i, (tuple, list)) for i in idx]):
all_plot_groups = pd.unique([tt for t in idx for tt in t]).tolist()
actual_groups_given = sum([len(i) for i in idx])
if actual_groups_given > len(all_plot_groups):
err0 = 'Groups are repeated across tuples,'
err1 = ' or a tuple has repeated groups in it.'
err2 = ' Please remove any duplicates and try again.'
raise ValueError(err0 + err1 + err2)
else: # mix of string and tuple?
err = 'There seems to be a problem with the idx you'
'entered--{}.'.format(idx)
raise ValueError(err)
# Having parsed the idx, check if it is a kosher paired plot,
# if so stated.
if paired is True:
all_idx_lengths = [len(t) for t in self.__idx]
if (np.array(all_idx_lengths) != 2).any():
err1 = "`is_paired` is True, but some idx "
err2 = "in {} does not consist only of two groups.".format(idx)
raise ValueError(err1 + err2)
# Determine the type of data: wide or long.
if x is None and y is not None:
err = 'You have only specified `y`. Please also specify `x`.'
raise ValueError(err)
elif y is None and x is not None:
err = 'You have only specified `x`. Please also specify `y`.'
raise ValueError(err)
# Identify the type of data that was passed in.
elif x is not None and y is not None:
# Assume we have a long dataset.
# check both x and y are column names in data.
if x not in data_in.columns:
err = '{0} is not a column in `data`. Please check.'.format(x)
raise IndexError(err)
if y not in data_in.columns:
err = '{0} is not a column in `data`. Please check.'.format(y)
raise IndexError(err)
# check y is numeric.
if not np.issubdtype(data_in[y].dtype, np.number):
err = '{0} is a column in `data`, but it is not numeric.'.format(y)
raise ValueError(err)
# check all the idx can be found in data_in[x]
for g in all_plot_groups:
if g not in data_in[x].unique():
err0 = '"{0}" is not a group in the column `{1}`.'.format(g, x)
err1 = " Please check `idx` and try again."
raise IndexError(err0 + err1)
# Select only rows where the value in the `x` column
# is found in `idx`.
plot_data = data_in[data_in.loc[:, x].isin(all_plot_groups)].copy()
# plot_data.drop("index", inplace=True, axis=1)
# Assign attributes
self.__x = x
self.__y = y
self.__xvar = x
self.__yvar = y
elif x is None and y is None:
# Assume we have a wide dataset.
# Assign attributes appropriately.
self.__x = None
self.__y = None
self.__xvar = "group"
self.__yvar = "value"
# First, check we have all columns in the dataset.
for g in all_plot_groups:
if g not in data_in.columns:
err0 = '"{0}" is not a column in `data`.'.format(g)
err1 = " Please check `idx` and try again."
raise IndexError(err0 + err1)
set_all_columns = set(data_in.columns.tolist())
set_all_plot_groups = set(all_plot_groups)
id_vars = set_all_columns.difference(set_all_plot_groups)
plot_data = pd.melt(data_in,
id_vars=id_vars,
value_vars=all_plot_groups,
value_name=self.__yvar,
var_name=self.__xvar)
# Added in v0.2.7.
# remove any NA rows.
plot_data.dropna(axis=0, how='any', subset=[self.__yvar], inplace=True)
# Lines 131 to 140 added in v0.2.3.
# Fixes a bug that jammed up when the xvar column was already
# a pandas Categorical. Now we check for this and act appropriately.
if isinstance(plot_data[self.__xvar].dtype,
pd.CategoricalDtype) is True:
plot_data[self.__xvar].cat.remove_unused_categories(inplace=True)
plot_data[self.__xvar].cat.reorder_categories(all_plot_groups,
ordered=True,
inplace=True)
else:
plot_data.loc[:, self.__xvar] = pd.Categorical(plot_data[self.__xvar],
categories=all_plot_groups,
ordered=True)
# # The line below was added in v0.2.4, removed in v0.2.5.
# plot_data.dropna(inplace=True)
self.__plot_data = plot_data
self.__all_plot_groups = all_plot_groups
# Sanity check that all idxs are paired, if so desired.
if paired is True:
if id_col is None:
err = "`id_col` must be specified if `is_paired` is set to True."
raise IndexError(err)
elif id_col not in plot_data.columns:
err = "{} is not a column in `data`. ".format(id_col)
raise IndexError(err)
EffectSizeDataFrame_kwargs = dict(ci=ci, is_paired=paired,
random_seed=random_seed,
resamples=resamples)
self.mean_diff = EffectSizeDataFrame(self, "mean_diff",
**EffectSizeDataFrame_kwargs)
self.median_diff = EffectSizeDataFrame(self, "median_diff",
**EffectSizeDataFrame_kwargs)
self.cohens_d = EffectSizeDataFrame(self, "cohens_d",
**EffectSizeDataFrame_kwargs)
self.hedges_g = EffectSizeDataFrame(self, "hedges_g",
**EffectSizeDataFrame_kwargs)
if paired is False:
self.cliffs_delta = EffectSizeDataFrame(self, "cliffs_delta",
**EffectSizeDataFrame_kwargs)
else:
self.cliffs_delta = "The data is paired; Cliff's delta is therefore undefined."
def __repr__(self):
from .__init__ import __version__
import datetime as dt
import numpy as np
from .misc_tools import print_greeting
if self.__is_paired:
es = "Paired e"
else:
es = "E"
greeting_header = print_greeting()
s1 = "{}ffect size(s) ".format(es)
s2 = "with {}% confidence intervals will be computed for:".format(self.__ci)
desc_line = s1 + s2
out = [greeting_header + "\n\n" + desc_line]
comparisons = []
for j, current_tuple in enumerate(self.__idx):
control_name = current_tuple[0]
for ix, test_name in enumerate(current_tuple[1:]):
comparisons.append("{} minus {}".format(test_name, control_name))
for j, g in enumerate(comparisons):
out.append("{}. {}".format(j+1, g))
resamples_line1 = "\n{} resamples ".format(self.__resamples)
resamples_line2 = "will be used to generate the effect size bootstraps."
out.append(resamples_line1 + resamples_line2)
return "\n".join(out)
# def __variable_name(self):
# return [k for k,v in locals().items() if v is self]
#
# @property
# def variable_name(self):
# return self.__variable_name()
@property
def data(self):
"""
Returns the pandas DataFrame that was passed to `dabest.load()`.
"""
return self.__data
@property
def idx(self):
"""
Returns the order of categories that was passed to `dabest.load()`.
"""
return self.__idx
@property
def is_paired(self):
"""
Returns True if the dataset was declared as paired to `dabest.load()`.
"""
return self.__is_paired
@property
def id_col(self):
"""
Returns the ic column declared to `dabest.load()`.
"""
return self.__id_col
@property
def ci(self):
"""
The width of the desired confidence interval.
"""
return self.__ci
@property
def resamples(self):
"""
The number of resamples used to generate the bootstrap.
"""
return self.__resamples
@property
def random_seed(self):
"""
The number used to initialise the numpy random seed generator, ie.
`seed_value` from `numpy.random.seed(seed_value)` is returned.
"""
return self.__random_seed
@property
def x(self):
"""
Returns the x column that was passed to `dabest.load()`, if any.
"""
return self.__x
@property
def y(self):
"""
Returns the y column that was passed to `dabest.load()`, if any.
"""
return self.__y
@property
def _xvar(self):
"""
Returns the xvar in dabest.plot_data.
"""
return self.__xvar
@property
def _yvar(self):
"""
Returns the yvar in dabest.plot_data.
"""
return self.__yvar
@property
def _plot_data(self):
"""
Returns the pandas DataFrame used to produce the estimation stats/plots.
"""
return self.__plot_data
@property
def _all_plot_groups(self):
"""
Returns the all plot groups, as indicated via the `idx` keyword.
"""
return self.__all_plot_groups
class TwoGroupsEffectSize(object):
"""
A class to compute and store the results of bootstrapped
mean differences between two groups.
"""
def __init__(self, control, test, effect_size,
is_paired=False, ci=95,
resamples=5000, random_seed=12345):
"""
Compute the effect size between two groups.
Parameters
----------
control : array-like
test : array-like
These should be numerical iterables.
effect_size : string.
Any one of the following are accepted inputs:
'mean_diff', 'median_diff', 'cohens_d', 'hedges_g', or 'cliffs_delta'
is_paired : boolean, default False
resamples : int, default 5000
The number of bootstrap resamples to be taken.
ci : float, default 95
The confidence interval width. The default of 95 produces 95%
confidence intervals.
random_seed : int, default 12345
`random_seed` is used to seed the random number generator during
bootstrap resampling. This ensures that the confidence intervals
reported are replicable.
Returns
-------
A :py:class:`TwoGroupEffectSize` object.
difference : float
The effect size of the difference between the control and the test.
effect_size : string
The type of effect size reported.
is_paired : boolean
Whether or not the difference is paired (ie. repeated measures).
ci : float
Returns the width of the confidence interval, in percent.
alpha : float
Returns the significance level of the statistical test as a float
between 0 and 1.
resamples : int
The number of resamples performed during the bootstrap procedure.
bootstraps : nmupy ndarray
The generated bootstraps of the effect size.
random_seed : int
The number used to initialise the numpy random seed generator, ie.
`seed_value` from `numpy.random.seed(seed_value)` is returned.
bca_low, bca_high : float
The bias-corrected and accelerated confidence interval lower limit
and upper limits, respectively.
pct_low, pct_high : float
The percentile confidence interval lower limit and upper limits,
respectively.
Examples
--------
>>> import numpy as np
>>> import scipy as sp
>>> import dabest
>>> np.random.seed(12345)
>>> control = sp.stats.norm.rvs(loc=0, size=30)
>>> test = sp.stats.norm.rvs(loc=0.5, size=30)
>>> effsize = dabest.TwoGroupsEffectSize(control, test, "mean_diff")
>>> effsize
The unpaired mean difference is -0.253 [95%CI -0.782, 0.241]
5000 bootstrap samples. The confidence interval is bias-corrected
and accelerated.
>>> effsize.to_dict()
{'alpha': 0.05,
'bca_high': 0.2413346581369784,
'bca_interval_idx': (109, 4858),
'bca_low': -0.7818088458343655,
'bootstraps': array([-1.09875628, -1.08840014, -1.08258695, ..., 0.66675324,
0.75814087, 0.80848265]),
'ci': 95,
'difference': -0.25315417702752846,
'effect_size': 'mean difference',
'is_paired': False,
'pct_high': 0.25135646125431527,
'pct_interval_idx': (125, 4875),
'pct_low': -0.763588353717278,
'pvalue_brunner_munzel': nan,
'pvalue_kruskal': nan,
'pvalue_mann_whitney': 0.2600723060808019,
'pvalue_paired_students_t': nan,
'pvalue_students_t': 0.34743913903372836,
'pvalue_welch': 0.3474493875548965,
'pvalue_wilcoxon': nan,
'random_seed': 12345,
'resamples': 5000,
'statistic_brunner_munzel': nan,
'statistic_kruskal': nan,
'statistic_mann_whitney': 406.0,
'statistic_paired_students_t': nan,
'statistic_students_t': 0.9472545159069105,
'statistic_welch': 0.9472545159069105,
'statistic_wilcoxon': nan}
"""
from numpy import array, isnan, isinf
from numpy import sort as npsort
from numpy.random import choice, seed
import scipy.stats as spstats
import lqrt
# import statsmodels.stats.power as power
from string import Template
import warnings
from ._stats_tools import confint_2group_diff as ci2g
from ._stats_tools import effsize as es
self.__EFFECT_SIZE_DICT = {"mean_diff" : "mean difference",
"median_diff" : "median difference",
"cohens_d" : "Cohen's d",
"hedges_g" : "Hedges' g",
"cliffs_delta" : "Cliff's delta"}
kosher_es = [a for a in self.__EFFECT_SIZE_DICT.keys()]
if effect_size not in kosher_es:
err1 = "The effect size '{}'".format(effect_size)
err2 = "is not one of {}".format(kosher_es)
raise ValueError(" ".join([err1, err2]))
if effect_size == "cliffs_delta" and is_paired is True:
err1 = "`paired` is True; therefore Cliff's delta is not defined."
raise ValueError(err1)
# Convert to numpy arrays for speed.
# NaNs are automatically dropped.
control = array(control)
test = array(test)
control = control[~isnan(control)]
test = test[~isnan(test)]
self.__effect_size = effect_size
self.__control = control
self.__test = test
self.__is_paired = is_paired
self.__resamples = resamples
self.__random_seed = random_seed
self.__ci = ci
self.__alpha = ci2g._compute_alpha_from_ci(ci)
self.__difference = es.two_group_difference(
control, test, is_paired, effect_size)
self.__jackknives = ci2g.compute_meandiff_jackknife(
control, test, is_paired, effect_size)
self.__acceleration_value = ci2g._calc_accel(self.__jackknives)
bootstraps = ci2g.compute_bootstrapped_diff(
control, test, is_paired, effect_size,
resamples, random_seed)
self.__bootstraps = npsort(bootstraps)
# Added in v0.2.6.
# Raises a UserWarning if there are any infiinities in the bootstraps.
num_infinities = len(self.__bootstraps[isinf(self.__bootstraps)])
if num_infinities > 0:
warn_msg = "There are {} bootstrap(s) that are not defined. "\
"This is likely due to smaple sample sizes. "\
"The values in a bootstrap for a group will be more likely "\
"to be all equal, with a resulting variance of zero. "\
"The computation of Cohen's d and Hedges' g thus "\
"involved a division by zero. "
warnings.warn(warn_msg.format(num_infinities),
category=UserWarning)
self.__bias_correction = ci2g.compute_meandiff_bias_correction(
self.__bootstraps, self.__difference)
# Compute BCa intervals.
bca_idx_low, bca_idx_high = ci2g.compute_interval_limits(
self.__bias_correction, self.__acceleration_value,
self.__resamples, ci)
self.__bca_interval_idx = (bca_idx_low, bca_idx_high)
if ~isnan(bca_idx_low) and ~isnan(bca_idx_high):
self.__bca_low = self.__bootstraps[bca_idx_low]
self.__bca_high = self.__bootstraps[bca_idx_high]
err1 = "The $lim_type limit of the interval"
err2 = "was in the $loc 10 values."
err3 = "The result should be considered unstable."
err_temp = Template(" ".join([err1, err2, err3]))
if bca_idx_low <= 10:
warnings.warn(err_temp.substitute(lim_type="lower",
loc="bottom"),
stacklevel=1)
if bca_idx_high >= resamples-9:
warnings.warn(err_temp.substitute(lim_type="upper",
loc="top"),
stacklevel=1)
else:
err1 = "The $lim_type limit of the BCa interval cannot be computed."
err2 = "It is set to the effect size itself."
err3 = "All bootstrap values were likely all the same."
err_temp = Template(" ".join([err1, err2, err3]))
if isnan(bca_idx_low):
self.__bca_low = self.__difference
warnings.warn(err_temp.substitute(lim_type="lower"),
stacklevel=0)
if isnan(bca_idx_high):
self.__bca_high = self.__difference
warnings.warn(err_temp.substitute(lim_type="upper"),
stacklevel=0)
# Compute percentile intervals.
pct_idx_low = int((self.__alpha/2) * resamples)
pct_idx_high = int((1-(self.__alpha/2)) * resamples)
self.__pct_interval_idx = (pct_idx_low, pct_idx_high)
self.__pct_low = self.__bootstraps[pct_idx_low]
self.__pct_high = self.__bootstraps[pct_idx_high]
# Perform statistical tests.
if is_paired is True:
# Wilcoxon, a non-parametric version of the paired T-test.
wilcoxon = spstats.wilcoxon(control, test)
self.__pvalue_wilcoxon = wilcoxon.pvalue
self.__statistic_wilcoxon = wilcoxon.statistic
if effect_size != "median_diff":
# Paired Student's t-test.
paired_t = spstats.ttest_rel(control, test, nan_policy='omit')
self.__pvalue_paired_students_t = paired_t.pvalue
self.__statistic_paired_students_t = paired_t.statistic
standardized_es = es.cohens_d(control, test, is_paired=True)
# self.__power = power.tt_solve_power(standardized_es,
# len(control),
# alpha=self.__alpha)
elif effect_size == "cliffs_delta":
# Let's go with Brunner-Munzel!
brunner_munzel = spstats.brunnermunzel(control, test,
nan_policy='omit')
self.__pvalue_brunner_munzel = brunner_munzel.pvalue
self.__statistic_brunner_munzel = brunner_munzel.statistic
elif effect_size == "median_diff":
# According to scipy's documentation of the function,
# "The Kruskal-Wallis H-test tests the null hypothesis
# that the population median of all of the groups are equal."
kruskal = spstats.kruskal(control, test, nan_policy='omit')
self.__pvalue_kruskal = kruskal.pvalue
self.__statistic_kruskal = kruskal.statistic
# self.__power = np.nan
else: # for mean difference, Cohen's d, and Hedges' g.
# Welch's t-test, assumes normality of distributions,
# but does not assume equal variances.
welch = spstats.ttest_ind(control, test, equal_var=False,
nan_policy='omit')
self.__pvalue_welch = welch.pvalue
self.__statistic_welch = welch.statistic
# Student's t-test, assumes normality of distributions,
# as well as assumption of equal variances.
students_t = spstats.ttest_ind(control, test, equal_var=True,
nan_policy='omit')
self.__pvalue_students_t = students_t.pvalue
self.__statistic_students_t = students_t.statistic
# Mann-Whitney test: Non parametric,
# does not assume normality of distributions
try:
mann_whitney = spstats.mannwhitneyu(control, test,
alternative='two-sided')
self.__pvalue_mann_whitney = mann_whitney.pvalue
self.__statistic_mann_whitney = mann_whitney.statistic
except ValueError:
# Occurs when the control and test are exactly identical
# in terms of rank (eg. all zeros.)
pass
# Likelihood Q-Ratio test:
# Assumes a gross-error model of distributions
try:
if self.__is_paired:
lqrt_ratio = lqrt.lqrtest_rel(control, test)
else:
lqrt_ratio = lqrt.lqrtest_ind(control, test)
self.__pvalue_lqrt = lqrt_ratio.pvalue
self.__statistic_lqrt = lqrt_ratio.statistic
except ImportError:
# did not install necessary library to run robust lqrt statistic
# warnings.warn("To get")
pass
standardized_es = es.cohens_d(control, test, is_paired=False)
# self.__power = power.tt_ind_solve_power(standardized_es,
# len(control),
# alpha=self.__alpha,
# ratio=len(test)/len(control)
# )
def __repr__(self, show_resample_count=True, define_pval=True, sigfig=3):
UNPAIRED_ES_TO_TEST = {"mean_diff" : "Mann-Whitney",
"median_diff" : "Kruskal",
"cohens_d" : "Mann-Whitney",
"hedges_g" : "Mann-Whitney",
"cliffs_delta" : "Brunner-Munzel"}
TEST_TO_PVAL_ATTR = {"Mann-Whitney" : "pvalue_mann_whitney",
"Kruskal" : "pvalue_kruskal",
"Brunner-Munzel" : "pvalue_brunner_munzel",
"Wilcoxon" : "pvalue_wilcoxon"}
PAIRED_STATUS = {True: 'paired', False: 'unpaired'}
first_line = {"is_paired": PAIRED_STATUS[self.__is_paired],
"es" : self.__EFFECT_SIZE_DICT[self.__effect_size]}
out1 = "The {is_paired} {es} ".format(**first_line)
base_string_fmt = "{:." + str(sigfig) + "}"
if "." in str(self.__ci):
ci_width = base_string_fmt.format(self.__ci)
else:
ci_width = str(self.__ci)
ci_out = {"es" : base_string_fmt.format(self.__difference),
"ci" : ci_width,
"bca_low" : base_string_fmt.format(self.__bca_low),
"bca_high" : base_string_fmt.format(self.__bca_high)}
out2 = "is {es} [{ci}%CI {bca_low}, {bca_high}].".format(**ci_out)
out = out1 + out2
if self.__is_paired:
stats_test = "Wilcoxon"
else:
stats_test = UNPAIRED_ES_TO_TEST[self.__effect_size]
pval_rounded = base_string_fmt.format(getattr(self,
TEST_TO_PVAL_ATTR[stats_test])
)
pvalue = "The two-sided p-value of the {} test is {}.".format(stats_test,
pval_rounded)
bs1 = "{} bootstrap samples were taken; ".format(self.__resamples)
bs2 = "the confidence interval is bias-corrected and accelerated."
bs = bs1 + bs2
defined = "The p-value(s) reported are the likelihood(s) of observing the " + \
"effect size(s),\nif the null hypothesis of zero difference is true."
if show_resample_count and define_pval:
return "{}\n{}\n\n{}\n{}".format(out, pvalue, bs, defined)
elif show_resample_count is False and define_pval is True:
return "{}\n{}\n\n{}".format(out, pvalue, defined)
elif show_resample_count is True and define_pval is False:
return "{}\n{}\n\n{}".format(out, pvalue, bs)
else:
return "{}\n{}".format(out, pvalue)
def to_dict(self):
"""
Returns the attributes of the `dabest.TwoGroupEffectSize` object as a
dictionary.
"""
# Only get public (user-facing) attributes.
attrs = [a for a in dir(self)
if not a.startswith(("_", "to_dict"))]
out = {}
for a in attrs:
out[a] = getattr(self, a)
return out
@property
def difference(self):
"""
Returns the difference between the control and the test.
"""
return self.__difference
@property
def effect_size(self):
"""
Returns the type of effect size reported.
"""
return self.__EFFECT_SIZE_DICT[self.__effect_size]
@property
def is_paired(self):
return self.__is_paired
@property
def ci(self):
"""
Returns the width of the confidence interval, in percent.
"""
return self.__ci
@property
def alpha(self):
"""
Returns the significance level of the statistical test as a float
between 0 and 1.
"""
return self.__alpha
@property
def resamples(self):
"""
The number of resamples performed during the bootstrap procedure.
"""
return self.__resamples
@property
def bootstraps(self):
"""
The generated bootstraps of the effect size.
"""
return self.__bootstraps
@property
def random_seed(self):
"""
The number used to initialise the numpy random seed generator, ie.
`seed_value` from `numpy.random.seed(seed_value)` is returned.
"""
return self.__random_seed
@property
def bca_interval_idx(self):
return self.__bca_interval_idx
@property
def bca_low(self):
"""
The bias-corrected and accelerated confidence interval lower limit.
"""
return self.__bca_low
@property
def bca_high(self):
"""
The bias-corrected and accelerated confidence interval upper limit.
"""
return self.__bca_high
@property
def pct_interval_idx(self):
return self.__pct_interval_idx
@property
def pct_low(self):
"""
The percentile confidence interval lower limit.
"""
return self.__pct_low
@property
def pct_high(self):
"""
The percentile confidence interval lower limit.
"""
return self.__pct_high
@property
def pvalue_brunner_munzel(self):
from numpy import nan as npnan
try:
return self.__pvalue_brunner_munzel
except AttributeError:
return npnan
@property
def statistic_brunner_munzel(self):
from numpy import nan as npnan
try:
return self.__statistic_brunner_munzel
except AttributeError:
return npnan
@property
def pvalue_wilcoxon(self):
from numpy import nan as npnan
try:
return self.__pvalue_wilcoxon
except AttributeError:
return npnan
@property
def statistic_wilcoxon(self):
from numpy import nan as npnan
try:
return self.__statistic_wilcoxon
except AttributeError:
return npnan
@property
def pvalue_lqrt(self):
from numpy import nan as npnan
try:
return self.__pvalue_lqrt
except AttributeError:
return npnan
@property
def statistic_lqrt(self):
from numpy import nan as npnan
try:
return self.__statistic_lqrt
except AttributeError:
return npnan
@property
def pvalue_paired_students_t(self):
from numpy import nan as npnan
try:
return self.__pvalue_paired_students_t
except AttributeError:
return npnan
@property
def statistic_paired_students_t(self):
from numpy import nan as npnan
try:
return self.__statistic_paired_students_t
except AttributeError:
return npnan
@property
def pvalue_kruskal(self):
from numpy import nan as npnan
try:
return self.__pvalue_kruskal
except AttributeError:
return npnan
@property
def statistic_kruskal(self):
from numpy import nan as npnan
try:
return self.__statistic_kruskal
except AttributeError:
return npnan
@property
def pvalue_welch(self):
from numpy import nan as npnan
try:
return self.__pvalue_welch
except AttributeError:
return npnan
@property
def statistic_welch(self):
from numpy import nan as npnan
try:
return self.__statistic_welch
except AttributeError:
return npnan
@property
def pvalue_students_t(self):
from numpy import nan as npnan
try:
return self.__pvalue_students_t
except AttributeError:
return npnan
@property
def statistic_students_t(self):
from numpy import nan as npnan
try:
return self.__statistic_students_t
except AttributeError:
return npnan
@property
def pvalue_mann_whitney(self):
from numpy import nan as npnan
try:
return self.__pvalue_mann_whitney
except AttributeError:
return npnan
@property
def statistic_mann_whitney(self):
from numpy import nan as npnan
try:
return self.__statistic_mann_whitney
except AttributeError:
return npnan