参考链接： Python中的numpy.flipud
numpy.argsort(), numpy.argmax(), numpy.argmin()用法
numpy.argsort(a, axis=-1, kind=’quicksort’, order=None)
功能：将矩阵a按照axis排序，并返回排序后的索引
参数：a为输入矩阵，axis为需要排序的维度，axis=0按列排序，axis=1按行排序
返回值：排序后的索引
# 一维向量
import numpy as np
a = np.array([1, 2, 3])
b = np.argsort(a)
print(b)
>> Out: [0 2 1]
# 二维向量，axis为默认值
import numpy as np
a = np.array([[1, 3, 2],[5, 7, 6]])
b = np.argsort(a)
print(b)
>> Out: [[0 2 1]
[0 2 1]]
# 二维向量，axis为0
import numpy as np
a = np.array([[1, 3, 2],[5, 7, 6]])
b = np.argsort(a, axis=0)
print(b)
>> Out: [[0 0 0]
[1 1 1]]
numpy.max(a, axis=-1, kind=’quicksort’, order=None)
功能：找到指定axis最大值，并返回最大值的索引
参数：a为输入矩阵，axis为寻找最大值的维度，axis=0按列寻找，axis=1按行寻找
返回值：最大值的索引
# 一维向量
import numpy as np
a = np.array([1, 2, 3])
b = np.argmax(a)
print(b)
>> Out: 2
# 二维向量，axis为默认值
import numpy as np
a = np.array([[1, 3, 2],[5, 7, 6]])
b = np.argmax(a)
print(b)
>> Out: 4
# 二维向量，axis为0
import numpy as np
a = np.array([[1, 3, 2],[5, 7, 6]])
b = np.argmax(a, axis=0)
print(b)
>> Out: [1 1 1]
# 二维向量，axis为1
import numpy as np
a = np.array([[1, 3, 2],[5, 7, 6]])
b = np.argmax(a, axis=1)
print(b)
>> Out: [1 1]
# 三维向量，axis为默认值
import numpy as np
a = np.array([[[1, 3, 2],[5, 7, 6]], [[4, 8, 6],[5, 7, 9]]])
b = np.argmax(a)
print(b)
>> Out: 11
# 三维向量，axis为0
import numpy as np
a = np.array([[[1, 3, 2],[5, 7, 6]], [[4, 8, 6],[5, 7, 9]]])
b = np.argmax(a, axis=0)
print(b)
>> Out: [[1 1 1]
[0 0 1]]
# 三维向量，axis为1
import numpy as np
a = np.array([[[1, 3, 2],[5, 7, 6]], [[4, 8, 6],[5, 7, 9]]])
b = np.argmax(a, axis=1)
print(b)
>> Out: [[1 1 1]
[1 0 1]]
numpy.min(a, axis=-1, kind=’quicksort’, order=None)
功能：找到指定axis最小值，并返回最小值的索引
参数：a为输入矩阵，axis为寻找最小值的维度，axis=0按列寻找，axis=1按行寻找
返回值：最小值的索引
