mxnet.np.average

average(a, axis=None, weights=None, returned=False, out=None)

Compute the weighted average along the specified axis.

Parameters

  • a (ndarray) – Array containing data to be averaged.

  • axis (None or int or tuple of ints, optional) – Axis or axes along which to average a. The default, axis=None, will average over all of the elements of the input array. If axis is negative it counts from the last to the first axis. New in version 1.7.0. If axis is a tuple of ints, averaging is performed on all of the axes specified in the tuple instead of a single axis or all the axes as before.

  • weights (ndarray, optional) – An array of weights associated with the values in a, must be the same dtype with a. Each value in a contributes to the average according to its associated weight. The weights array can either be 1-D (in which case its length must be the size of a along the given axis) or of the same shape as a. If weights=None, then all data in a are assumed to have a weight equal to one. The 1-D calculation is: avg = sum(a * weights) / sum(weights) The only constraint on weights is that sum(weights) must not be 0.

  • returned (bool, optional) – Default is False. If True, the tuple (average, sum_of_weights) is returned, otherwise only the average is returned. If weights=None, sum_of_weights is equivalent to the number of elements over which the average is taken.

  • out (ndarray, optional) – If provided, the calculation is done into this array.

Returns

retval, [sum_of_weights] – Return the average along the specified axis. When returned is True, return a tuple with the average as the first element and the sum of the weights as the second element. sum_of_weights is of the same type as retval. If a is integral, the result dtype will be current default dtype, When npx.is_np_default_dtype() returns False, default dtype is float32, When npx.is_np_default_dtype() returns True, default dtype is float64; otherwise it will be the same as dtype of a.

Return type

ndarray

Raises

  • MXNetError –

  • * When all weights along axis sum to zero. –

  • * When the length of 1D weights is not the same as the shape of a along axis. –

  • * When given 1D weights, the axis is not specified or is not int. –

  • * When the shape of weights and a differ, but weights are not 1D. –

Note

This function differs from the original numpy.average <https://numpy.org/devdocs/reference/generated/numpy.average.html>`_ in the following way(s):

  • Does not guarantee the same behavior with numpy when given float16 dtype and overflow happens

  • Does not support complex dtype

  • The dtypes of a and weights must be the same

  • Integral a results in float32 or float64 returned dtype:

    • When npx.is_np_default_dtype() returns False, default dtype is float32,

    • When npx.is_np_default_dtype() returns True, default dtype is float64;

Examples

>>> data = np.arange(1, 5)
>>> data
array([1., 2., 3., 4.])
>>> np.average(data)
array(2.5)
>>> np.average(np.arange(1, 11), weights=np.arange(10, 0, -1))
array(4.)
>>> data = np.arange(6).reshape((3,2))
>>> data
array([[0., 1.],
       [2., 3.],
       [4., 5.]])
>>> weights = np.array([0.25, 0.75])
array([0.25, 0.75])
>>> np.average(data, axis=1, weights=weights)
array([0.75, 2.75, 4.75])