um is selected over multiple
    axes, instead of a single axis or all the axes as before.
out : array_like, optional
    Alternative output array in which to place the result.  Must
    be of the same shape and buffer length as the expected output.
fill_value : scalar or None, optional
    Value used to fill in the masked values.
    If None, use the output of maximum_fill_value().
keepdims : bool, optional
    If this is set to True, the axes which are reduced are left
    in the result as dimensions with size one. With this option,
    the result will broadcast correctly against the array.

Returns
-------
amax : array_like
    New array holding the result.
    If ``out`` was specified, ``out`` is returned.

See Also
--------
ma.maximum_fill_value
    Returns the maximum filling value for a given datatype.

Examples
--------
>>> import numpy.ma as ma
>>> x = [[-1., 2.5], [4., -2.], [3., 0.]]
>>> mask = [[0, 0], [1, 0], [1, 0]]
>>> masked_x = ma.masked_array(x, mask)
>>> masked_x
masked_array(
  data=[[-1.0, 2.5],
        [--, -2.0],
        [--, 0.0]],
  mask=[[False, False],
        [ True, False],
        [ True, False]],
  fill_value=1e+20)
>>> ma.max(masked_x)
2.5
>>> ma.max(masked_x, axis=0)
masked_array(data=[-1.0, 2.5],
             mask=[False, False],
       fill_value=1e+20)
>>> ma.max(masked_x, axis=1, keepdims=True)
masked_array(
  data=[[2.5],
        [-2.0],
        [0.0]],
  mask=[[False],
        [False],
        [False]],
  fill_value=1e+20)
>>> mask = [[1, 1], [1, 1], [1, 1]]
>>> masked_x = ma.masked_array(x, mask)
>>> ma.max(masked_x, axis=1)
masked_array(data=[--, --, --],
             mask=[ True,  True,  True],
       fill_value=1e+20,
            dtype=float64)
r4