normtest 0.0.2

Check that your data follows, at least approximately, the Normal distribution.

normtest

This package has a series of tests used to check whether a set of sample data follows, at least approximately, the Normal distribution.

Available tests (17/11/2023)

• Filliben
• Ryan-Joiner

Install

pip install normtest

Usage

Each test has its own class and can be imported as follows:

from normtest import RyanJoiner
from normtest import Filliben

To perform the test, just instantiate the class and apply the fit method, passing the data set as a NumpyArray. For example:

import numpy as np
test = RyanJoiner()
x_data = np.array([6, 1, -4, 8, -2, 5, 0])
test.fit(x_data)

After the fit method is applied, the test object now has a series of attributes with the test results. The main attribute is test.normality, which contains the summarized results:

print(test.normality)
RyanJoiner(statistic=0.9844829186140105, critical=0.8977794003662074, p_value='p > 0.100', conclusion='Fail to reject H₀')

The test object also has methods for graphical visualization of results, such as the line_up method. See the documentation for details.

Each test has its individual module, and functions can be accessed through the modules. To import the module that contains all the RyanJoiner test functions, for example, use:

from normtest import ryan_joiner

This way, it is possible to generate graphs and obtain intermediate values from the test calculations. For example:

size = 7
pi = ryan_joiner._order_statistic(size)
print(pi)
[0.0862069  0.22413793 0.36206897 0.5        0.63793103 0.77586207
 0.9137931 ]

License

• BSD 3-Clause License