numpy库学习 向量 矩阵 均为有两个[[ ,而秩为1的数组只有一个[ np.array([[]]) 与np.array([])的区别

对于一维普通数组取shape 时只有第一项有值 举例 （4，）

对于二维普通数组取shape 时才会有完整的第二项值 举例 （4，1）

 

https://blog.csdn.net/zenghaitao0128/article/details/78300770

 

什么是秩

https://www.zhihu.com/question/21605094

 

自己实验

import numpy as np
X=np.array([[-0.60116],
 [-0.94159],
 [-0.74565],
 [0.89583]])
w = [0]
b=1
z= np.matmul(X,w) + b

print(X.shape)

W = np.zeros(X.shape[1]) 
print("W=",W)
print("W.TYPE=",type(W),"w.shape=",W.shape)

W2 = np.zeros(X.shape[0]) 
print("W2=",W2)
print("W2.TYPE=",type(W2),"w2.shape=",W2.shape)

W3=np.array([[0, 0, 0, 0]])
print("W3=",W3)
print("W3.TYPE=",type(W3),"w3.shape=",W3.shape) 

W4 = np.zeros(4) 
print("W4=",W4)
print("W4.TYPE=",type(W4),"w4.shape=",W4.shape)

W5 = np.array([[0,0,0,0]]) 
print("W5=",W5)
print("W5.TYPE=",type(W5),"w5.shape=",W5.shape)

W6 = np.array([0,0,0,0]) 
print("W6=",W6)
print("W6.TYPE=",type(W6),"w6.shape=",W6.shape)

(4, 1)
W= [0.]
W.TYPE= <class 'numpy.ndarray'> w.shape= (1,)
W2= [0. 0. 0. 0.]
W2.TYPE= <class 'numpy.ndarray'> w2.shape= (4,)
W3= [[0 0 0 0]]
W3.TYPE= <class 'numpy.ndarray'> w3.shape= (1, 4)
W4= [0. 0. 0. 0.]
W4.TYPE= <class 'numpy.ndarray'> w4.shape= (4,)
W5= [[0 0 0 0]]
W5.TYPE= <class 'numpy.ndarray'> w5.shape= (1, 4)
W6= [0 0 0 0]
W6.TYPE= <class 'numpy.ndarray'> w6.shape= (4,)