The dimension NumPy (dimension), shaft (axis), rank (rank) meaning (in-depth understanding highly recommend)
https://blog.csdn.net/Babyfatliang/article/details/87721282
- This article explains: axis dimension = = rank. And how to calculate, how to understand the calculation.
- You can get a glimpse of twelve by this picture.
The dimension of the array to be understood as a nested structure (briefly described)
https://blog.csdn.net/qq_42383041/article/details/86157702
a = np.array([[1,2,3],[1,2,3]])
print(np.sum(a,axis = 0))
>>>[2 4 6]
print(np.sum(a,axis = 1))
>>>[6 6]
print(np.sum(a))
>>>12
- The first [2,3] considered as x, the second [2,3] seen y, changes arr [x, y] which can be seen as a one-dimensional array in accordance with a so-dimensional I appreciated that the array x, y in this order on the axis 0, i.e. the axis 0 [x, y] = [[1,2,3], [1,2,3]].
- When Axis = 1, a first dimension (a first nesting) unchanged, i.e., the addition of a second dimension (a second nesting), that is [3,1 + 1 + 2 + 3 + 2] = [6,6].
- When no parameters axis, all the elements are added, namely 12.
- Dimensionality reduction method to calculate, if the dimensions down to two-dimensional, then it is for us to understand, simply means "replace, dimension reduction combat."
An array of supplemental index
https://blog.csdn.net/sinat_34072381/article/details/84448307
- Nested slightly different here, a first number represents the first array dimension, the second dimension representing a second number, the third number represents the third dimension, and so on.
- E.g. A [[0,1]] 0,1 axis represents the position of elements, the result array dimension is 1, and A [[[0,1]]] Equivalent
- Note that here, the index of the first layer brackets or square brackets, can be separated by a comma, with the increase in the partition of the index will increase the number of dimensions of the array results. Such as: d = b [[[0,0], [1,1]], [[1,3], [2,3]]] returns the intersection region.
