Skip to content

Commit f51eb68

Browse files
Jacobluke-Jacobluke-
committed
Update the dabest.load() method. Follow nbdev
1 parent 011d017 commit f51eb68

2 files changed

Lines changed: 2 additions & 422 deletions

File tree

‎dabest/_classes.py‎

Lines changed: 1 addition & 211 deletions
Original file line numberDiff line numberDiff line change
@@ -438,29 +438,7 @@ def __repr__(self):
438438
@property
439439
def mean_diff(self):
440440
"""
441-
Returns an :py:class:`EffectSizeDataFrame` for the mean difference, its confidence interval, and relevant statistics, for all comparisons as indicated via the `idx` and `paired` argument in `dabest.load()`.
442-
443-
Example
444-
-------
445-
>>> from scipy.stats import norm
446-
>>> import pandas as pd
447-
>>> import dabest
448-
>>> control = norm.rvs(loc=0, size=30, random_state=12345)
449-
>>> test = norm.rvs(loc=0.5, size=30, random_state=12345)
450-
>>> my_df = pd.DataFrame({"control": control,
451-
"test": test})
452-
>>> my_dabest_object = dabest.load(my_df, idx=("control", "test"))
453-
>>> my_dabest_object.mean_diff
454-
455-
Notes
456-
-----
457-
This is simply the mean of the control group subtracted from
458-
the mean of the test group.
459-
460-
.. math::
461-
\\text{Mean difference} = \\overline{x}_{Test} - \\overline{x}_{Control}
462-
463-
where :math:`\\overline{x}` is the mean for the group :math:`x`.
441+
Returns an :py:class:`EffectSizeDataFrame` for the mean difference, its confidence interval, and relevant statistics, for all comparisons as indicated via the `idx` and `paired` argument in `dabest.load()`
464442
465443
"""
466444
return self.__mean_diff
@@ -471,45 +449,6 @@ def median_diff(self):
471449
"""
472450
Returns an :py:class:`EffectSizeDataFrame` for the median difference, its confidence interval, and relevant statistics, for all comparisons as indicated via the `idx` and `paired` argument in `dabest.load()`.
473451
474-
475-
Example
476-
-------
477-
>>> from scipy.stats import norm
478-
>>> import pandas as pd
479-
>>> import dabest
480-
>>> control = norm.rvs(loc=0, size=30, random_state=12345)
481-
>>> test = norm.rvs(loc=0.5, size=30, random_state=12345)
482-
>>> my_df = pd.DataFrame({"control": control,
483-
"test": test})
484-
>>> my_dabest_object = dabest.load(my_df, idx=("control", "test"))
485-
>>> my_dabest_object.median_diff
486-
487-
488-
Notes
489-
-----
490-
This is the median difference between the control group and the test group.
491-
492-
If the comparison(s) are unpaired, median_diff is computed with the following equation:
493-
494-
.. math::
495-
\\text{Median difference} = \\widetilde{x}_{Test} - \\widetilde{x}_{Control}
496-
497-
where :math:`\\widetilde{x}` is the median for the group :math:`x`.
498-
499-
If the comparison(s) are paired, median_diff is computed with the following equation:
500-
501-
.. math::
502-
\\text{Median difference} = \\widetilde{x}_{Test - Control}
503-
504-
505-
Things to note
506-
--------------
507-
Using median difference as the statistic in bootstrapping may result in a biased estimate and cause problems with BCa confidence intervals. Consider using mean difference instead.
508-
509-
When plotting, consider using percentile confidence intervals instead of BCa confidence intervals by specifying `ci_type = 'percentile'` in .plot().
510-
511-
For detailed information, please refer to `Issue 129 <https://github.com/ACCLAB/DABEST-python/issues/129>`_.
512-
513452
"""
514453
return self.__median_diff
515454

@@ -518,59 +457,6 @@ def median_diff(self):
518457
def cohens_d(self):
519458
"""
520459
Returns an :py:class:`EffectSizeDataFrame` for the standardized mean difference Cohen's `d`, its confidence interval, and relevant statistics, for all comparisons as indicated via the `idx` and `paired` argument in `dabest.load()`.
521-
522-
Example
523-
-------
524-
>>> from scipy.stats import norm
525-
>>> import pandas as pd
526-
>>> import dabest
527-
>>> control = norm.rvs(loc=0, size=30, random_state=12345)
528-
>>> test = norm.rvs(loc=0.5, size=30, random_state=12345)
529-
>>> my_df = pd.DataFrame({"control": control,
530-
"test": test})
531-
>>> my_dabest_object = dabest.load(my_df, idx=("control", "test"))
532-
>>> my_dabest_object.cohens_d
533-
534-
Notes
535-
-----
536-
Cohen's `d` is simply the mean of the control group subtracted from
537-
the mean of the test group.
538-
539-
If `paired` is None, then the comparison(s) are unpaired;
540-
otherwise the comparison(s) are paired.
541-
542-
If the comparison(s) are unpaired, Cohen's `d` is computed with the following equation:
543-
544-
.. math::
545-
546-
d = \\frac{\\overline{x}_{Test} - \\overline{x}_{Control}} {\\text{pooled standard deviation}}
547-
548-
549-
For paired comparisons, Cohen's d is given by
550-
551-
.. math::
552-
d = \\frac{\\overline{x}_{Test} - \\overline{x}_{Control}} {\\text{average standard deviation}}
553-
554-
where :math:`\\overline{x}` is the mean of the respective group of observations, :math:`{Var}_{x}` denotes the variance of that group,
555-
556-
.. math::
557-
558-
\\text{pooled standard deviation} = \\sqrt{ \\frac{(n_{control} - 1) * {Var}_{control} + (n_{test} - 1) * {Var}_{test} } {n_{control} + n_{test} - 2} }
559-
560-
and
561-
562-
.. math::
563-
564-
\\text{average standard deviation} = \\sqrt{ \\frac{{Var}_{control} + {Var}_{test}} {2}}
565-
566-
The sample variance (and standard deviation) uses N-1 degrees of freedoms.
567-
This is an application of `Bessel's correction <https://en.wikipedia.org/wiki/Bessel%27s_correction>`_, and yields the unbiased
568-
sample variance.
569-
570-
References:
571-
https://en.wikipedia.org/wiki/Effect_size#Cohen's_d
572-
https://en.wikipedia.org/wiki/Bessel%27s_correction
573-
https://en.wikipedia.org/wiki/Standard_deviation#Corrected_sample_standard_deviation
574460
575461
"""
576462
return self.__cohens_d
@@ -581,37 +467,6 @@ def cohens_h(self):
581467
"""
582468
Returns an :py:class:`EffectSizeDataFrame` for the standardized mean difference Cohen's `h`, its confidence interval, and relevant statistics, for all comparisons as indicated via the `idx` and `directional` argument in `dabest.load()`.
583469
584-
Example
585-
-------
586-
>>> from scipy.stats import randint
587-
>>> import pandas as pd
588-
>>> import dabest
589-
>>> control = randint.rvs(0, 2, size=30, random_state=12345)
590-
>>> test = randint.rvs(0, 2, size=30, random_state=12345)
591-
>>> my_df = pd.DataFrame({"control": control,
592-
"test": test})
593-
>>> my_dabest_object = dabest.load(my_df, idx=("control", "test")
594-
>>> my_dabest_object.cohens_h
595-
596-
Notes
597-
-----
598-
Cohen's *h* uses the information of proportion in the control and test groups to calculate the distance between two proportions.
599-
It can be used to describe the difference between two proportions as "small", "medium", or "large".
600-
It can be used to determine if the difference between two proportions is "meaningful".
601-
602-
A directional Cohen's *h* is computed with the following equation:
603-
604-
.. math::
605-
h = 2 * \\arcsin{\\sqrt{proportion_{Test}}} - 2 * \\arcsin{\\sqrt{proportion_{Control}}}
606-
607-
For a non-directional Cohen's *h*, the equation is:
608-
609-
.. math::
610-
h = |2 * \\arcsin{\\sqrt{proportion_{Test}}} - 2 * \\arcsin{\\sqrt{proportion_{Control}}}|
611-
612-
References:
613-
https://en.wikipedia.org/wiki/Cohen%27s_h
614-
615470
"""
616471
return self.__cohens_h
617472

@@ -620,40 +475,6 @@ def cohens_h(self):
620475
def hedges_g(self):
621476
"""
622477
Returns an :py:class:`EffectSizeDataFrame` for the standardized mean difference Hedges' `g`, its confidence interval, and relevant statistics, for all comparisons as indicated via the `idx` and `paired` argument in `dabest.load()`.
623-
624-
625-
Example
626-
-------
627-
>>> from scipy.stats import norm
628-
>>> import pandas as pd
629-
>>> import dabest
630-
>>> control = norm.rvs(loc=0, size=30, random_state=12345)
631-
>>> test = norm.rvs(loc=0.5, size=30, random_state=12345)
632-
>>> my_df = pd.DataFrame({"control": control,
633-
"test": test})
634-
>>> my_dabest_object = dabest.load(my_df, idx=("control", "test"))
635-
>>> my_dabest_object.hedges_g
636-
637-
Notes
638-
-----
639-
640-
Hedges' `g` is :py:attr:`cohens_d` corrected for bias via multiplication with the following correction factor:
641-
642-
.. math::
643-
\\frac{ \\Gamma( \\frac{a} {2} )} {\\sqrt{ \\frac{a} {2} } \\times \\Gamma( \\frac{a - 1} {2} )}
644-
645-
where
646-
647-
.. math::
648-
a = {n}_{control} + {n}_{test} - 2
649-
650-
and :math:`\\Gamma(x)` is the `Gamma function <https://en.wikipedia.org/wiki/Gamma_function>`_.
651-
652-
653-
654-
References:
655-
https://en.wikipedia.org/wiki/Effect_size#Hedges'_g
656-
https://journals.sagepub.com/doi/10.3102/10769986006002107
657478
658479
"""
659480
return self.__hedges_g
@@ -663,37 +484,6 @@ def hedges_g(self):
663484
def cliffs_delta(self):
664485
"""
665486
Returns an :py:class:`EffectSizeDataFrame` for Cliff's delta, its confidence interval, and relevant statistics, for all comparisons as indicated via the `idx` and `paired` argument in `dabest.load()`.
666-
667-
668-
Example
669-
-------
670-
>>> from scipy.stats import norm
671-
>>> import pandas as pd
672-
>>> import dabest
673-
>>> control = norm.rvs(loc=0, size=30, random_state=12345)
674-
>>> test = norm.rvs(loc=0.5, size=30, random_state=12345)
675-
>>> my_df = pd.DataFrame({"control": control,
676-
"test": test})
677-
>>> my_dabest_object = dabest.load(my_df, idx=("control", "test"))
678-
>>> my_dabest_object.cliffs_delta
679-
680-
681-
Notes
682-
-----
683-
684-
Cliff's delta is a measure of ordinal dominance, ie. how often the values from the test sample are larger than values from the control sample.
685-
686-
.. math::
687-
\\text{Cliff's delta} = \\frac{\\#({x}_{test} > {x}_{control}) - \\#({x}_{test} < {x}_{control})} {{n}_{Test} \\times {n}_{Control}}
688-
689-
690-
where :math:`\\#` denotes the number of times a value from the test sample exceeds (or is lesser than) values in the control sample.
691-
692-
Cliff's delta ranges from -1 to 1; it can also be thought of as a measure of the degree of overlap between the two samples. An attractive aspect of this effect size is that it does not make an assumptions about the underlying distributions that the samples were drawn from.
693-
694-
References:
695-
https://en.wikipedia.org/wiki/Effect_size#Effect_size_for_ordinal_data
696-
https://psycnet.apa.org/record/1994-08169-001
697487
698488
"""
699489
return self.__cliffs_delta

0 commit comments

Comments
 (0)