Source code for pynance.learn.metrics

"""
.. Copyright (c) 2016 Marshall Farrier
   license http://opensource.org/licenses/MIT

Metrics (:mod:`pynance.learn.metrics`)
===============================================

.. currentmodule:: pynance.learn.metrics
"""

import numpy as np

[docs]def mse(predicted, actual): """ Mean squared error of predictions. .. versionadded:: 0.5.0 Parameters ---------- predicted : ndarray Predictions on which to measure error. May contain a single or multiple column but must match `actual` in shape. actual : ndarray Actual values against which to measure predictions. Returns ------- err : ndarray Mean squared error of predictions relative to actual values. """ diff = predicted - actual return np.average(diff * diff, axis=0)
[docs]def stderr(predicted, actual): """ Standard error of predictions. .. versionadded:: 0.5.0 Parameters ---------- predicted : ndarray Predictions on which to measure error. May contain a single or multiple column but must match `actual` in shape. actual : ndarray Actual values against which to measure predictions. Returns ------- err : ndarray Standard error of predictions relative to actual values. """ return np.sqrt(mse(predicted, actual))