mqr.inference.hyptest.

HypothesisTest

class mqr.inference.hyptest.HypothesisTest(description: str, alternative: str, method: str, sample_stat: str, sample_stat_target: str, sample_stat_value: float64, stat: float64, pvalue: float64, null: str = None, alt: str = None)

Result of an hypothesis test.

Attributes:

descriptionstr

Description of the statistic on which the test is performed.

alternative{‘two-sided’, ‘less’, ‘greater’}

Sense of the alternative hypothesis.

methodstr

Name of the test method, eg. ‘Kruskall-Wallace’.

sample_statstr

Name of the statistic from the sample on which the test is performed. For example, when performing an hypothesis test on the mean of a sample, sample_stat is the mean. This is independent of the method.

sample_stat_targetstr

Hypothesised value of the sample statistic. For example, when performing an hypothesis test on the mean of a sample, sample_stat_target is the hypothesised mean that appears in the null-hypothesis. Expressed as a string because some hypothesis tests target non-numeric properties, like distributions.

sample_stat_valuefloat

Actual value of the sample statistic calculated from the sample. Use None when the sample_stat_target is non-numeric.

statfloat

Test statistic. For example, in an hypothesis test on the mean of a sample, stat might be the score on a student’s-t distribution.

pvaluefloat

p-value associated with the test statistic.

nullNone

Automatically generated. A string representation of the null-hypothesis.

altNone

Automatically generated. A string representation of the alternative-hypothesis.

Notes

This class is iterable, and iterates over the test statistic and p-value (in that order).

Examples

All hypothesis tests return a type constructed like this.

>>> import mqr
>>> hyptest = mqr.inference.hyptest.HypothesisTest(
>>>     description='Example mean test',
>>>     alternative='less',
>>>     method='Example students-t',
>>>     sample_stat='mean(x)',
>>>     sample_stat_target='1.6',
>>>     sample_stat_value=1.4,
>>>     stat=-2.828427124746192,
>>>     pvalue=0.02371032779215975)
>>> hyptest
HypothesisTest(
    description='Example mean test',
    alternative='less',
    method='Example students-t',
    sample_stat='mean(x)',
    sample_stat_target='1.6',
    sample_stat_value=1.4,
    stat=-2.828427124746192,
    pvalue=0.02371032779215975,
    null='mean(x) == 1.6',
    alt='mean(x) < 1.6')

Iterate over the test statistic and p-value:

>>> stat, pvalue = hyptest
>>> stat, pvalue
(-2.828427124746192, 0.02371032779215975)

When displayed in a jupyter notebook, the example here will be shown as an HTML table:

Hypothesis Test Example mean test
method	Example students-t
H0	mean(x) == 1.6
H1	mean(x) < 1.6
statistic	-2.828427124746192
p-value	0.02371032779215975