I need to make a code generation from Python to Matlab, and I'm new to Python.
Although there are some websites which give definitions of what numpy.bincount and what numpy.argmax are, they don't give a simple example which is understandable for new beginners.
There is such a statement in a Python algorithm:
numpy.argmax(numpy.bincount(dlabel))
dlabel is a matrix. What does this function exactly do? This calculation is very important for me, and I don't want to make rush because of misunderstanding of this simple code statement.
Could you show me example for numpy.bincount, and numpy.argmax on matrices differently, and briefly?
numpy.argmax(numpy.bincount(dlabel)) returns the most common value found in dlabel.
To break it down, np.bincount() will return the count of each value in an array of non-negative integers and return an array with the count at the appropriate index (Python arrays are indexed from 0). For example:
>>> np.bincount(np.array([1,2,3,2,4,1,1,0]))
array([1, 3, 2, 1, 1], dtype=int32)
# So '0' appears once, '1' appears three times, ...
It also works for int arrays that don't start at zero
>>> np.bincount(np.array([3, 3, 3, 3, 3, 4]))
array([0, 0, 0, 5, 1], dtype=int32)
# So '3' appears 5 times, '4' appears three times, ...
np.argmax() returns the index of the (first) maximum value in an array.
>>> np.argmax([1, 3, 2, 1, 1])
1
# maximum value is '3', with index '1'
So combining the two functions, we can find that 1 appears most often in [1,2,3,2,4,1,1,0] (it appears 3 times).
>>> np.argmax(np.bincount([1,2,3,2,4,1,1,0]))
1