Numpy.array percentile of the matrix (scaled)

Disclaimer: This article is a blogger original article, shall not be reproduced without the bloggers allowed. https://blog.csdn.net/a19990412/article/details/90745159

Brief

Use Needless to say, very common

• In the case of DataFrame, but the operation is different. Note contrast.
• Pandas.DataFrame rows required percentage (logarithmic scale)

problem

• Suppose data A

>>> A
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19]])

The ratio of the number becomes how to operate? Direct line dividing the sum?

>>> A / A.sum(axis=1)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: operands could not be broadcast together with shapes (4,5) (4,)

Solution

• numpy Generally, only the corresponding bit of the operation, or the value (in fact, understood as a vector of length will be more accurate) and vector operations.
• Thus it requires np.newaxis operable to shape the same into numpy.array

>>> A / A.sum(axis=1)[:, np.newaxis]
array([[0.        , 0.1       , 0.2       , 0.3       , 0.4       ],
       [0.14285714, 0.17142857, 0.2       , 0.22857143, 0.25714286],
       [0.16666667, 0.18333333, 0.2       , 0.21666667, 0.23333333],
       [0.17647059, 0.18823529, 0.2       , 0.21176471, 0.22352941]])

• A.sum(axis=1)[:, np.newaxis]What is? It is to copy a lot of times it's vector

>>> A.sum(axis=1)
array([10, 35, 60, 85])
>>> A.sum(axis=1)[:, np.newaxis]
array([[10],
       [35],
       [60],
       [85]])