Python numpy.nanquantile()

numpy.nanquantile(arr, q, axis = None) :计算给定数据（数组元素）沿指定轴的q th quantile，忽略nan值。当人们处理正态分布时，量化在统计学中起着非常重要的作用。

在上图中，Q2是正态分布数据的中位数。Q3-Q2代表给定数据集的**四分位数范围。

参数 :

arr : [array_like]输入阵列。

q :量化值。

axis : [int or tuples of int]轴，我们想沿着这个轴计算量化值。axis = 0表示沿列工作， axis = 1表示沿行工作。

输出: [ndarray, optional]不同的数组，我们想把结果放在其中。该数组必须具有与预期输出相同的尺寸。

Results : 数组的第3个四分位数（如果轴为零，则为标量值）或数组中沿指定轴的四分位值，忽略nan值。

代码#1：

# Python Program illustrating
# numpy.nanquantile() method 
import numpy as np
     
# 1D array
arr = [20, 2, 7, np.nan, 34]
print("arr : ", arr)
 
print("\n-Q1 quantile of arr : ", np.quantile(arr, .50))
print("Q2 - quantile of arr : ", np.quantile(arr, .25))
print("Q3 - quantile of arr : ", np.quantile(arr, .75))
 
print("\nQ1 - nanquantile of arr : ", np.nanquantile(arr, .50))
print("Q2 - nanquantile of arr : ", np.nanquantile(arr, .25))
print("Q3 - nanquantile of arr : ", np.nanquantile(arr, .75))

输出 :

arr : [20, 2, 7, nan, 34]

Q1 - quantile of arr : nan
Q2 - quantile of arr : nan
Q3 - quantile of arr : nan

Q1 - nanquantile of arr : 13.5
Q2 - nanquantile of arr : 5.75
Q3 - nanquantile of arr : 23.5

代码 #2:

# Python Program illustrating
# numpy.nanquantile() method
 
import numpy as np
 
# 2D array
arr = [[14, np.nan, 12, 33, 44],
       [15, np.nan, 27, 8, 19],
       [23, 2, np.nan, 1, 4, ]]
print("\narr : \n", arr)
     
# quantile of the flattened array
print("\nQ2 quantile of arr, axis = None : ", np.quantile(arr, .50))
print("\nQ2 quantile of arr, axis = None : ", np.nanquantile(arr, .50))
print("0th quantile of arr, axis = None : ", np.nanquantile(arr, 0))

输出:

arr : 
[[14, nan, 12, 33, 44], [15, nan, 27, 8, 19], [23, 2, nan, 1, 4]]

Q2 quantile of arr, axis = None : nan
Q2 quantile of arr, axis = None : 14.5
0th quantile of arr, axis = None : 1.0

代码 #3:

# Python Program illustrating
# numpy.nanquantile() method
import numpy as np
 
# 2D array
arr = [[14, np.nan, 12, 33, 44],
    [15, np.nan, 27, 8, 19],
    [23, 2, np.nan, 1, 4, ]]
print("\narr : \n", arr)
         
# quantile along the axis = 0
print("\nQ2 quantile of arr, axis = 0 : ", np.nanquantile(arr, .50, axis = 0))
print("0th quantile of arr, axis = 0 : ", np.nanquantile(arr, 0, axis = 0))
 
# quantile along the axis = 1
print("\nQ2 quantile of arr, axis = 1 : ", np.nanquantile(arr, .50, axis = 1))
print("0th quantile of arr, axis = 1 : ", np.nanquantile(arr, 0, axis = 1))
 
print("\nQ2 quantile of arr, axis = 1 : \n",
  np.nanquantile(arr, .50, axis = 1, keepdims = True))
print("\n0th quantile of arr, axis = 1 : \n",
    np.nanquantile(arr, 0, axis = 1, keepdims = True))

输出:

arr : 
[[14, nan, 12, 33, 44], [15, nan, 27, 8, 19], [23, 2, nan, 1, 4]]

Q2 quantile of arr, axis = 0 : [15.  2. 19.5  8.  19. ]
0th quantile of arr, axis = 0 : [14. 2. 12.  1.  4.]

Q2 quantile of arr, axis = 1 : [23.5 17.   3. ]
0th quantile of arr, axis = 1 : [12.  8.  1.]

Q2 quantile of arr, axis = 1 : 
[[23.5]
[17. ]
[ 3. ]]

0th quantile of arr, axis = 1 : 
[[12.]
[ 8.]
[ 1.]]