Numpy库基础___五

Numpy数据存取

•NumPy的随机数函数

a = np.random.rand(1,2,3)
print(a)
#[[[0.03339719 0.72784732 0.47527802]
#  [0.6456671  0.65639799 0.01300073]]]

a = np.random.randn(1,2,3)
print(a)
#[[[ 0.59115211 -0.40289048  1.34532466]
#  [-0.04616715 -0.64066822 -1.09722129]]]

a = np.random.randint(100,200,(3,4))
print(a)
#[[161 131 187 134]
# [156 114 104 180]
# [182 163 158 121]]

#随机数种子，10是给定的种子值
np.random.seed(10)
a = np.random.randint(100,200,(3,4))
print(a)
#[[109 115 164 128]
# [189 193 129 108]
# [173 100 140 136]]

a = np.random.randint(100,200,(3,4))
print(a)
#[[184 199 152 144]
# [173 171 179 144]
# [133 105 197 143]]

np.random.shuffle(a)
print(a)
#[[173 171 179 144]
# [133 105 197 143]
# [184 199 152 144]]

b = np.random.permutation(a)
#[[173 171 179 144]
# [133 105 197 143]
# [184 199 152 144]]
print(b)
#[[133 105 197 143]
# [173 171 179 144]
# [184 199 152 144]]

a = np.random.randint(100,200,(8,))
print(a)
#[131 195 130 165 177 107 197 132]

b = np.random.choice(a,(3,2))
print(b)
#[[195 107]
# [177 197]
# [130 107]]

b = np.random.choice(a,(3,2),replace=False)
#[[107 130]
# [197 132]
# [195 131]]

#加权，元素出现次数越多，被抽取的概率越高
b = np.random.choice(a,(3,2),p=a/np.sum(a)) 
print(b)
#[[197 130]
# [131 130]
# [131 130]]

u = np.random.uniform(0,10,(3,4))
print(u)
#[[7.49328353 4.35990777 8.19266316 5.02229727]
# [2.21122875 9.61785352 9.90294149 2.44401573]
# [3.88367203 9.22037768 7.87306998 2.00241521]]

u = np.random.normal(10,5,(3,4))
print(u)
#[[13.44007699 10.5502136  14.79616224 -2.17381553]
# [10.42238979 10.12351539  2.8561042  16.78322252]
# [11.90679396  6.75343566  8.01259211 14.96874378]]

u = np.random.poission(2,(3,4))
print(u)
#[[4 0 1 2]
# [2 2 3 2]
# [0 0 2 3]]

---

•NumPy的统计函数

a = np.arange(15).reshape(3,5)
print(a)
#[[ 0  1  2  3  4]
# [ 5  6  7  8  9]
# [10 11 12 13 14]]
print(np.sum(a))
#105
print(np.sum(a,axis=0))
#[15 18 21 24 27]
print(np.sum(a,axis=1))
#[10 35 60]

print(np.mean(a))
#7.0
print(np.mean(a,axis=0))
#[5. 6. 7. 8. 9.]
print(np.mean(a,axis=1))
#[ 2.  7. 12.]

print(np.average(a))
#7.0
print(np.average(a,axis=0,weights=[1,2,3]))
#[ 6.66666667  7.66666667  8.66666667  9.66666667 10.66666667]

a = np.arange(12).reshape(3,4)
print(a)
#[[ 0  1  2  3]
# [ 4  5  6  7]
# [ 8  9 10 11]]

print(np.min(a))
#0

print(np.max(a))
#11

print(np.argmin(a))
#0

print(np.argmax(a))
#11

print(np.unravel_index(10,(4,3)))
#(3,1)

print(np.unravel_index(np.argmax(a),(4,3)))
#(3,2)

print(np.ptp(a))
#11

print(np.median(a))
#5.5

---

•NumPy的梯度函数

• np.gradient(f):计算数组f中元素的梯度,当f为多维时，返回每个维度梯度

　　　梯度：连续值之间的变化率，即斜率

　　　X坐标轴连续三个x坐标对应的Y轴值：a,b,c其中b的梯度时(c-a)/2

a = np.random.randint(0,20,(5,))
print(a)
#[ 2 10 11 14 12]

print(np.gradient(a))
#[ 8.   4.5  2.   0.5 -2. ]