Question

我正在从规范化数组中访问X和Y值。但是当我使用for循环将一个方程应用于规范化数组的每个bin时，想要用0排除值。我收到0值的错误。

     normed = array([[ 0.0003,  0.0026,  0.0149,  0.0288,  0.0328,  0.0174,  0.0036,
     0.0004,  0.    ,  0.    ],
   [ 0.0001,  0.0024,  0.0103,  0.0244,  0.0321,  0.0211,  0.0064,
     0.0006,  0.    ,  0.    ],
   [ 0.0003,  0.0012,  0.0072,  0.0235,  0.0325,  0.0246,  0.0079,
     0.0012,  0.    ,  0.    ],
   [ 0.    ,  0.0008,  0.0065,  0.0182,  0.0346,  0.0274,  0.0124,
     0.002 ,  0.0004,  0.    ],
   [ 0.    ,  0.0005,  0.0033,  0.0155,  0.0302,  0.0327,  0.0153,
     0.0032,  0.0003,  0.    ],
   [ 0.    ,  0.0002,  0.0025,  0.0121,  0.0283,  0.0342,  0.0184,
     0.0054,  0.0004,  0.    ],
   [ 0.    ,  0.0001,  0.0016,  0.008 ,  0.0253,  0.0318,  0.0219,
     0.0084,  0.001 ,  0.0002],
   [ 0.    ,  0.    ,  0.0008,  0.0056,  0.0198,  0.0316,  0.0296,
     0.0108,  0.0018,  0.0002],
   [ 0.    ,  0.0001,  0.0007,  0.0035,  0.0169,  0.0329,  0.0318,
     0.0163,  0.0039,  0.0006],
   [ 0.    ,  0.    ,  0.0001,  0.0022,  0.0107,  0.0253,  0.0329,
     0.0171,  0.004 ,  0.0011]])

    Px = normed.sum(axis=0)
    array([ 0.0007,  0.0079,  0.0479,  0.1418,  0.2632,  0.279 ,  0.1802,
    0.0654,  0.0118,  0.0021])

    def funct():
    CEnt = 0
        for i in range(0,10):
             if i > 0:
                 for j in range(0,10):
                     if j > 0:
                     CEnt += normed[i][j]*np.log2(Px/(normed[i][j]))
        return CEnt

当我打印funct（）时，我得到了错误：[nan nan nan nan nan nan nan nan nan] RuntimeWarning：在true_divide中遇到的除以零 RuntimeWarning：在乘法中遇到无效值

如何使用CEnt方程编写for循环来遍历数组中的100个bin，同时排除0值？非常感谢....

Answer 1

您将索引与零进行比较而不是实际值。以下代码将忽略零。

import numpy as np
from numpy import array

normed = array([[ 0.0003,  0.0026,  0.0149,  0.0288,  0.0328,  0.0174,  0.0036,
     0.0004,  0.    ,  0.    ],
   [ 0.0001,  0.0024,  0.0103,  0.0244,  0.0321,  0.0211,  0.0064,
     0.0006,  0.    ,  0.    ],
   [ 0.0003,  0.0012,  0.0072,  0.0235,  0.0325,  0.0246,  0.0079,
     0.0012,  0.    ,  0.    ],
   [ 0.    ,  0.0008,  0.0065,  0.0182,  0.0346,  0.0274,  0.0124,
     0.002 ,  0.0004,  0.    ],
   [ 0.    ,  0.0005,  0.0033,  0.0155,  0.0302,  0.0327,  0.0153,
     0.0032,  0.0003,  0.    ],
   [ 0.    ,  0.0002,  0.0025,  0.0121,  0.0283,  0.0342,  0.0184,
     0.0054,  0.0004,  0.    ],
   [ 0.    ,  0.0001,  0.0016,  0.008 ,  0.0253,  0.0318,  0.0219,
     0.0084,  0.001 ,  0.0002],
   [ 0.    ,  0.    ,  0.0008,  0.0056,  0.0198,  0.0316,  0.0296,
     0.0108,  0.0018,  0.0002],
   [ 0.    ,  0.0001,  0.0007,  0.0035,  0.0169,  0.0329,  0.0318,
     0.0163,  0.0039,  0.0006],
   [ 0.    ,  0.    ,  0.0001,  0.0022,  0.0107,  0.0253,  0.0329,
     0.0171,  0.004 ,  0.0011]])

Px = normed.sum(axis=0)

def funct():
    CEnt = 0
    for line in normed:
         for value in line:
             if value > 0:
                CEnt += value*np.log2(Px/(value))
    return CEnt

print(funct())

循环（对于这个10x10阵列）与此相同：

def funct():
    CEnt = 0
    for i in range(10):
         for j in range(10):
             if normed[i][j] > 0:
                CEnt += normed[i][j]*np.log2(Px/(normed[i][j]))
    return CEnt

访问规范化数组并在for循环中排除0值

1 个答案: