numpy高阶应用
随机数
| 类型 |
说明 |
| seed |
确定随机数生成器的种子 |
| permutation |
返回一个序列的随机排列或返回一个随机排列的返回 |
| shuffle |
对一个序列就地随机乱序 |
| rand |
产生均匀分布的样本值 |
| randint |
从给定的上下限范围内随机选取整数 |
| randn |
产生正态分布（平均值为0，标准差为1） |
| binomial |
产生二项分布的样本值 |
| normal |
产生正态（高斯）分布的样本值 |
| beta |
产生Beta分布的样本值 |
| chisquare |
产生卡方分布的样本值 |
| gamma |
产Gamma分布的样本值 |
| uniform |
产生在[0, 1]中均匀分布的样本值 |
import numpy as np
arr = np.random.normal(size = 10)
print(arr)
[ 0.20551433 -0.47894623 0.1723548 -2.52035522 0.90416095 -0.0748014
0.2370496 -1.08262676 -0.29973029 0.48172872]
数组重塑
将一维数组转化为二维数组
arr = np.arange(15)
print(arr.reshape(3,5))
[[ 0 1 2 3 4]
[ 5 6 7 8 9]
[10 11 12 13 14]]
获取维度信息并应用
other_arr = np.arange(15).reshape(5,3)
print(other_arr)
print(arr.reshape(other_arr.shape))
[[ 0 1 2]
[ 3 4 5]
[ 6 7 8]
[ 9 10 11]
[12 13 14]]
[[ 0 1 2]
[ 3 4 5]
[ 6 7 8]
[ 9 10 11]
[12 13 14]]
数组拉平
arr = arr.reshape(other_arr.shape)
print(arr)
print(arr.ravel())
print(arr.ravel().reshape(other_arr.shape))
[[ 0 1 2]
[ 3 4 5]
[ 6 7 8]
[ 9 10 11]
[12 13 14]]
[ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14]
[[ 0 1 2]
[ 3 4 5]
[ 6 7 8]
[ 9 10 11]
[12 13 14]]
数组连接
| 类型 |
说明 |
| concatenate |
最一般化的连接，沿一条轴连接一组数组 |
| vstack,row_stack |
以面向行的方式对数组进行堆叠（沿轴0） |
| hstack, |
以面向行的方式对数组进行堆叠（沿轴1） |
| column_stack |
类似于hstack，但是会先将一维数组转换为二维列向量。 |
| dstack |
以面向"深度"的方式对数组进行堆叠（沿轴2） |
| split |
沿指定轴在指定的位置拆分数组 |
| hsplit,vsplit,dsplit |
split的便捷化函数，分别沿着轴0、轴1和轴2进行拆分。 |
arr1 = np.array([[1,2,3,4],[5,6,7,8]])
arr2 = np.array([[9,10,11,12],[13,14,15,16]])
print(np.concatenate([arr1,arr2],axis = 0)) #按行，向下连接
print(np.concatenate([arr1,arr2],axis = 1)) #按列，向右连接
[[ 1 2 3 4]
[ 5 6 7 8]
[ 9 10 11 12]
[13 14 15 16]]
[[ 1 2 3 4 9 10 11 12]
[ 5 6 7 8 13 14 15 16]]
数组拆分
arr = np.random.randn(5,5)
print(arr)
a,b,c = np.split(arr,[1,3],axis = 0)
print(a)
print(b)
print(c)
#help(np.split)
[[ 0.65808685 -1.87367158 -0.44571392 0.20153713 1.07337974]
[ 0.56055164 0.81087505 -0.92973517 -0.54649625 -0.251764 ]
[-1.05954976 0.89089206 -1.47122268 -0.88814653 -0.35668902]
[ 1.16943732 -0.30638432 -1.09877266 0.33839846 1.43588489]
[-0.94883322 0.81643331 -0.3468645 -0.63871218 -1.26688086]]
[[ 0.65808685 -1.87367158 -0.44571392 0.20153713 1.07337974]]
[[ 0.56055164 0.81087505 -0.92973517 -0.54649625 -0.251764 ]
[-1.05954976 0.89089206 -1.47122268 -0.88814653 -0.35668902]]
[[ 1.16943732 -0.30638432 -1.09877266 0.33839846 1.43588489]
[-0.94883322 0.81643331 -0.3468645 -0.63871218 -1.26688086]]
堆叠水平和垂直
print(arr)
arr1 = np.array([[1, 2, 3], [4, 5, 6]])
arr2 = np.array([[7, 8, 9], [10, 11, 12]])
print(np.vstack([arr1,arr2]))
print(np.hstack([arr1,arr2]))
[[ 0.65808685 -1.87367158 -0.44571392 0.20153713 1.07337974]
[ 0.56055164 0.81087505 -0.92973517 -0.54649625 -0.251764 ]
[-1.05954976 0.89089206 -1.47122268 -0.88814653 -0.35668902]
[ 1.16943732 -0.30638432 -1.09877266 0.33839846 1.43588489]
[-0.94883322 0.81643331 -0.3468645 -0.63871218 -1.26688086]]
[[ 1 2 3]
[ 4 5 6]
[ 7 8 9]
[10 11 12]]
[[ 1 2 3 7 8 9]
[ 4 5 6 10 11 12]]
堆叠辅助类 r_ c_
print(np.r_[arr1,arr2])
print(np.c_[arr1,arr2])
[[ 1 2 3]
[ 4 5 6]
[ 7 8 9]
[10 11 12]]
[[ 1 2 3 7 8 9]
[ 4 5 6 10 11 12]]
元素的重复操作
arr = np.random.randn(2,2)
print(arr.repeat(2,axis = 0))
print(arr.repeat(2,axis = 1))
[[-0.44388113 0.13987511]
[-0.44388113 0.13987511]
[ 1.42439852 2.53537756]
[ 1.42439852 2.53537756]]
[[-0.44388113 -0.44388113 0.13987511 0.13987511]
[ 1.42439852 1.42439852 2.53537756 2.53537756]]
tile
print(np.tile(arr,2))
print(np.tile(arr,[2,1]))
[[-0.44388113 0.13987511 -0.44388113 0.13987511]
[ 1.42439852 2.53537756 1.42439852 2.53537756]]
[[-0.44388113 0.13987511]
[ 1.42439852 2.53537756]
[-0.44388113 0.13987511]
[ 1.42439852 2.53537756]]
花式索引
arr = np.arange(10) * 100
inds = [7, 1, 2, 6]
print(arr[inds])
[700 100 200 600]
等效的索引
print(arr.take(inds))
arr.put(inds,[0,0,0,0])
print(arr)
arr = np.random.rand(3,4)
inds = [2, 1, 2, 2]
print(arr.take(inds,axis = 1)) #按列查找
[700 100 200 600]
[ 0 0 0 300 400 500 0 0 800 900]
[[ 0.6876503 0.08996047 0.6876503 0.6876503 ]
[ 0.60100624 0.18482635 0.60100624 0.60100624]
[ 0.77159319 0.04447536 0.77159319 0.77159319]]
