arkouda.scipy
=============

.. py:module:: arkouda.scipy


Submodules
----------

.. toctree::
   :maxdepth: 1

   /autoapi/arkouda/scipy/special/index
   /autoapi/arkouda/scipy/stats/index


Classes
-------

.. autoapisummary::

   arkouda.scipy.Power_divergenceResult


Functions
---------

.. autoapisummary::

   arkouda.scipy.chisquare
   arkouda.scipy.power_divergence


Module Contents
---------------

.. py:class:: Power_divergenceResult

   Bases: :py:obj:`Power_divergenceResult`


   The results of a power divergence statistical test.

   .. attribute:: statistic



      :type: numpy.float64

   .. attribute:: pvalue



      :type: numpy.float64


.. py:function:: chisquare(f_obs, f_exp=None, ddof=0)

   Computes the chi square statistic and p-value.

   :param f_obs: The observed frequency.
   :type f_obs: pdarray
   :param f_exp: The expected frequency.
   :type f_exp: pdarray, default = None
   :param ddof: The delta degrees of freedom.
   :type ddof: int

   :rtype: arkouda.akstats.Power_divergenceResult

   .. rubric:: Examples

   >>> import arkouda as ak
   >>> ak.connect()
   >>> from arkouda.stats import chisquare
   >>> chisquare(ak.array([10, 20, 30, 10]), ak.array([10, 30, 20, 10]))
   Power_divergenceResult(statistic=8.333333333333334, pvalue=0.03960235520756414)

   .. seealso:: :obj:`scipy.stats.chisquare`, :obj:`arkouda.akstats.power_divergence`

   .. rubric:: References

   [1] “Chi-squared test”, https://en.wikipedia.org/wiki/Chi-squared_test

   [2] "scipy.stats.chisquare",
   https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.chisquare.html


.. py:function:: power_divergence(f_obs, f_exp=None, ddof=0, lambda_=None)

   Computes the power divergence statistic and p-value.

   :param f_obs: The observed frequency.
   :type f_obs: pdarray
   :param f_exp: The expected frequency.
   :type f_exp: pdarray, default = None
   :param ddof: The delta degrees of freedom.
   :type ddof: int
   :param lambda_: The power in the Cressie-Read power divergence statistic.
                   Allowed values: "pearson", "log-likelihood", "freeman-tukey", "mod-log-likelihood",
                   "neyman", "cressie-read"

                   Powers correspond as follows:

                   "pearson": 1

                   "log-likelihood": 0

                   "freeman-tukey": -0.5

                   "mod-log-likelihood": -1

                   "neyman": -2

                   "cressie-read": 2 / 3
   :type lambda_: string, default = "pearson"

   :rtype: arkouda.akstats.Power_divergenceResult

   .. rubric:: Examples

   >>> import arkouda as ak
   >>> ak.connect()
   >>> from arkouda.stats import power_divergence
   >>> x = ak.array([10, 20, 30, 10])
   >>> y = ak.array([10, 30, 20, 10])
   >>> power_divergence(x, y, lambda_="pearson")
   Power_divergenceResult(statistic=8.333333333333334, pvalue=0.03960235520756414)
   >>> power_divergence(x, y, lambda_="log-likelihood")
   Power_divergenceResult(statistic=8.109302162163285, pvalue=0.04380595350226197)

   .. seealso:: :obj:`scipy.stats.power_divergence`, :obj:`arkouda.akstats.chisquare`

   .. rubric:: Notes

   This is a modified version of scipy.stats.power_divergence [2]
   in order to scale using arkouda pdarrays.

   .. rubric:: References

   [1] "scipy.stats.power_divergence",
   https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.power_divergence.html

   [2] Scipy contributors (2024) scipy (Version v1.12.0) [Source code].
   https://github.com/scipy/scipy