cs231n-knn-note

import numpy as np

下面验证一下 np.concatenate 的用法

a = np.array([[1,2],[3,4]])
b = np.array([[5,6],[7,8]])
x = []
x.append(a)
x.append(b)
x

[array([[1, 2],
        [3, 4]]), array([[5, 6],
        [7, 8]])]

发现不管是列表还是元组都是可以的

x1 = np.concatenate([a,b])
x1

array([[1, 2],
       [3, 4],
       [5, 6],
       [7, 8]])

x2 = np.concatenate((a,b))
x2

array([[1, 2],
       [3, 4],
       [5, 6],
       [7, 8]])

def my_sum(a, order=0):
    sum = 0
    for i in a:
        sum += i
    return sum

my_sum([1,2,3], 0)

6

my_sum((1,2,3), 0)

6

因为在 for 循环中不仅列表可以循环，元组，字符只要是可迭代的都可以循环

而在下面的my_sum1 函数的定义中，我们是以可变参数来定义的

def my_sum1(*a, order=0):
    sum = 0;
    for i in a:
        sum += i
    return sum

my_sum1(1,2,3)

6

my_sum1([1,2,3])

---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

<ipython-input-20-0a7a90b048ef> in <module>()
----> 1 my_sum1([1,2,3])


<ipython-input-18-1de405cf845d> in my_sum1(order, *a)
      2     sum = 0;
      3     for i in a:
----> 4         sum += i
      5     return sum


TypeError: unsupported operand type(s) for +=: 'int' and 'list'

my_sum1(*[1,2,3])

6

如果用 np.concatenate(a1,a2…)会报错

x3 = np.concatenate(a,b)
x3

---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

<ipython-input-22-9c75f38c761f> in <module>()
----> 1 x3 = np.concatenate(a,b)
      2 x3


TypeError: only integer scalar arrays can be converted to a scalar index

综上 np.concatenate 的调用只能 np.concatenate((a1,a2,…)) 或者 np.concatenate([a1,a2,…])

比较异同

a = np.array([1,2,3]) # 可以把它当成列向量
b = np.array([4,5,6,7])
np.concatenate((a,b))

array([1, 2, 3, 4, 5, 6, 7])

np.concatenate((a,b), axis=0)

array([1, 2, 3, 4, 5, 6, 7])

np.concatenate((a,b), axis=None)

array([1, 2, 3, 4, 5, 6, 7])

np.concatenate((a,b), axis=1)

---------------------------------------------------------------------------

ValueError                                Traceback (most recent call last)

<ipython-input-79-2a8f7c46ccd2> in <module>()
----> 1 np.concatenate((a,b), axis=1)


ValueError: all the input array dimensions except for the concatenation axis must match exactly

无法对一维数组进行 axis = 1 上的操作，因为一维数组只有一个纵向的维度

a = np.array([1,2,3]) 
b = np.array([4,5,6])
np.concatenate((a,b), axis=1)

---------------------------------------------------------------------------

AxisError                                 Traceback (most recent call last)

<ipython-input-80-2528feefaf7f> in <module>()
      1 a = np.array([1,2,3])
      2 b = np.array([4,5,6])
----> 3 np.concatenate((a,b), axis=1)


AxisError: axis 1 is out of bounds for array of dimension 1

a = np.array([[1,2,3]])
b = np.array([[4,5,6,7]])
np.concatenate((a,b))

---------------------------------------------------------------------------

ValueError                                Traceback (most recent call last)

<ipython-input-71-4b1407c41995> in <module>()
      1 a = np.array([[1,2,3]])
      2 b = np.array([[4,5,6,7]])
----> 3 np.concatenate((a,b))


ValueError: all the input array dimensions except for the concatenation axis must match exactly

a = np.array([[1],[2],[3]])
b = np.array([[4],[5],[7]])
np.concatenate((a,b),axis=1)

array([[1, 4],
       [2, 5],
       [3, 7]])

np.concatenate((a,b), axis=1)

array([[1, 2, 3, 4, 5, 6, 7]])

a = np.array([[1,2],[5,4]])
b = np.array([1,2])
a + b

array([[2, 4],
       [6, 6]])

a = np.array([[1,2],[5,4]])
b = np.array([[1,2]])
a + b

array([[2, 4],
       [6, 6]])

np.flatnonzero

np.flatnonzero(a) 返回非0数组的下标的平铺形式

比较：
a.ravel() 返回数组的平铺形式

np.flatnonzero(a)

array([0, 1, 2, 3])

a.ravel()

array([1, 2, 3, 4])

a = np.array([[1,2,3]])
b = np.array([[3,4,5]])
np.sum((a-b)**2)

12

a-b

array([[-2, -2, -2]])

np.argsort

返回数组从小到大的序号

a = np.array([[3,1,2],[4,7,1]])
np.argsort(a)

array([[1, 2, 0],
       [2, 0, 1]])

np 中找到出现最多的元素

a = np.array([1,2,3,3,4,4,4,5])
x = np.bincount(a)
x

array([0, 1, 1, 2, 3, 1])

np.argmax(x)

4

np.sum

a = np.array([[1,2,3],[4,5,6]])
np.sum(a,axis=0)

array([5, 7, 9])

np.sum(a,axis=1)

array([ 6, 15])

np.sum(a)

21

会发现不管axis 是0还是1得出的都是(n,)这样的数组

切片

a = np.array([1,2,3])
a.shape

(3,)

[1] + [2]

[1, 2]

[1,2,3][0:1]

[1]

[1,2,3][1:2]

[2]

[1,2,3][0:0]

[]

x1 = []
x2 = []
x1.append(a)
x2.append(a)

比较异同

x1[0:]

[array([1, 2, 3])]

x1[0]

array([1, 2, 3])

x1[0:] + x2[0:]

[array([1, 2, 3]), array([1, 2, 3])]