.. _features:

Features
********

At its core *effectus*

  * identifies :meth:`whether <.Effects.pareto>` a pareto distribution is present. [#f1]_
  * identifies the **most relevant cause-effect relationship** (:term:`Vital Few`). [#f1]_
  * identifies the :meth:`Rule 50/5 <effectus.Effects.the_rule>` or a variant of it
    (related but not identical to :term:`Useful Many`). [#f2]_
  * determines :meth:`causes <effectus.Effects.attain_causes>` and
    :meth:`effects <effectus.Effects.attain_effects>` for any given share
    of effects and causes, respectively.
  * identifies the **separating value** for the :term:`Vital Few`. [#f3]_
  * assesses the change af the :term:`arithmetic mean <Arithmetic Mean>`
    and the :term:`standard deviation <Standard Deviation>` for the
    split of the :term:`Vital Few` and the :term:`Useful Many`.

Builtins
========

  * :ref:`Intervals <intervals>` module: Allows for any interval of causes or effects.
  * :ref:`Intersection <intersections>` module: Automagically
    intersects pareto distributions of result and effort attributes into
    ABCD classification and allows to
    :ref:`add and subtract <bitwise_operators>` the groups easily.
  * TBD: :ref:`Nested Pareto distributions <nested>`
  * :ref:`Throughput Accounting <throughput>`
  * 30 select :ref:`examples <example-data>` from real life
  * :ref:`Excel & CSV interfaces <interfaces>`
  * :ref:`Python API <api>`: Fully documented.

Performance
===========

*effectus* performs well for any number of values (parallel computation; MacBook Air, SSD):

    ==========  ================
    values      ms for 1k values
    ==========  ================
    0-1000      71-3.3
    1000-10000  3.6-2.9
    > 10000     3.3-1.4
    ==========  ================

.. [#f1] Representation of :class:`effectus.Effects`
.. [#f2] :meth:`effectus.Effects.the_rule`.
.. [#f3] :meth:`effectus.Effects.separate_causes` & :meth:`effectus.Effects.separate_effects`.