Module ``pyqt_fit.bootstrap``
=============================

.. automodule:: pyqt_fit.bootstrap

.. currentmodule:: pyqt_fit.bootstrap

Bootstrap Shuffling Methods
---------------------------

.. autofunction:: bootstrap_residuals

.. autofunction:: bootstrap_regression

Main Boostrap Functions
-----------------------

.. autofunction:: bootstrap

.. class:: BootstrapResult(y_fit, y_est, y_eval, CIs, shuffled_xs, shuffled_ys, full_results)

  .. note::

    This is a class created with :py:func:`pyqt_fit.utils.namedtuple`.

  .. py:attribute:: y_fit

            Estimator object, fitted on the original data
            :type: fun(xs) -> ys

  .. py:attribute:: y_est

            Y estimated on xdata
            :type: ndarray

  .. py:attribute:: eval_points

            Points on which the confidence interval are evaluated

  .. py:attribute:: y_eval

            Y estimated on eval_points

  .. py:attribute:: CIs_val

            Tuple containing the list of percentiles extracted (i.e. this is a copy of
            the ``CIs`` argument of the bootstrap function.

  .. py:attribute:: CIs

            List of confidence intervals. The first element is for the estimated values
            on ``eval_points``. The others are for the extra attributes specified in
            ``extra_attrs``. Each array is a 3-dimensional array (Q,2,N), where
            Q is the number of confidence interval (e.g. the length of ``CIs_val``)
            and N is the number of data points. Values (x,0,y) give the lower bounds
            and (x,1,y) the upper bounds of the confidence intervals.

  .. py:attribute:: shuffled_xs

            if full_results is True, the shuffled x's used for the bootstrapping

  .. py:attribute:: shuffled_ys

            if full_results is True, the shuffled y's used for the bootstrapping

  .. py:attribute:: full_results

            if full_results is True, the estimated y's for each shuffled_ys