In [1]:

import numpy as np

In [2]:

python内置对象list是一维数组，而numpy提供了n维数组及其运算，其数据类型是ndarray，运算法则依然遵循list方式，对于多维数组、矢量以及矩阵运算非常方便。

In [4]:

#将list转换为ndarray
np.array([[1,2], [3,4]])

Out[4]:

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

In [6]:

#生成一维序列，再转换为ndarray
np.arange(25).reshape(5,5)

Out[6]:

array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])

In [13]:

#也可以随机生成多维数组
np.random.rand(3,4)
#randn生成的数据符合正太分布
np.random.randn(4,4)

Out[13]:

array([[-0.17294533,  0.57664369,  0.90030628, -1.50687522],
       [ 0.44645627, -0.007905  , -0.52326875, -0.60296276],
       [ 0.90004056,  0.76595149, -0.51941328,  0.40380163],
       [-0.5386898 ,  1.68159334, -0.07916927, -1.28261209]])

In [72]:

#区间采样：[start, end]
np.random.uniform(10, 30, 100)

Out[72]:

array([12.0695157 , 15.28852086, 28.16887606, 23.78457455, 19.43228899,
       19.4356459 , 29.21877117, 23.19956387, 19.72328742, 18.5866164 ,
       23.65236814, 22.90528204, 14.61641437, 19.12403056, 18.71734194,
       21.86738754, 12.61084045, 16.48857546, 10.13327166, 29.18626043,
       13.52597009, 25.3757283 , 15.02568918, 17.58892082, 29.02929713,
       19.31601472, 18.70991889, 27.60091717, 17.90855395, 22.32056116,
       29.66164178, 15.80899303, 13.42051403, 21.63006552, 27.3575489 ,
       19.15324054, 11.6461633 , 15.2618994 , 17.25711308, 11.15194372,
       24.7998248 , 20.68732859, 19.61350829, 13.52628516, 20.93910938,
       18.22947301, 22.55097367, 19.83017006, 28.02903298, 21.88747298,
       25.62843102, 25.9743397 , 28.41009972, 19.96665695, 12.90643362,
       10.53082376, 25.83417207, 14.47090988, 12.30504558, 25.61192155,
       29.43102168, 17.16377112, 25.98836525, 10.56868229, 19.95139715,
       10.80447219, 18.29030416, 16.87592569, 25.22108043, 25.97517119,
       13.60687878, 15.40081725, 18.56386783, 29.06878897, 25.60131042,
       18.75444808, 13.6712598 , 20.53491322, 20.62849974, 24.44350913,
       22.15870654, 25.66238923, 14.30747784, 20.37187352, 24.98596286,
       18.57617349, 10.26463449, 11.84077318, 23.86112526, 11.08118911,
       26.51167248, 19.01538728, 18.25704808, 10.36148126, 10.69855404,
       18.29660878, 14.1800761 , 27.21586491, 20.89498324, 13.51931409])

In [16]:

#全为1的数组
np.ones((2,2))

Out[16]:

array([[1., 1.],
       [1., 1.]])

In [17]:

#全为0的数组
np.zeros((2,2))

Out[17]:

array([[0., 0.],
       [0., 0.]])

In [24]:

#对角矩阵
np.eye(5, 5, 1)

Out[24]:

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

In [28]:

#获取对角元素
np.diag(np.arange(25).reshape(5,5))

Out[28]:

array([ 0,  6, 12, 18, 24])

In [36]:

#numpy提供了直接加载cvs格式数据的接口。delimiter是指定分隔符，usecols指定选择数据的列
filename = 'ex1data2.txt'
np.loadtxt(filename, delimiter=',', usecols=(0, 2))

Out[36]:

array([[  2104., 399900.],
       [  1600., 329900.],
       [  2400., 369000.],
       [  1416., 232000.],
       [  3000., 539900.],
       [  1985., 299900.],
       [  1534., 314900.],
       [  1427., 198999.],
       [  1380., 212000.],
       [  1494., 242500.],
       [  1940., 239999.],
       [  2000., 347000.],
       [  1890., 329999.],
       [  4478., 699900.],
       [  1268., 259900.],
       [  2300., 449900.],
       [  1320., 299900.],
       [  1236., 199900.],
       [  2609., 499998.],
       [  3031., 599000.],
       [  1767., 252900.],
       [  1888., 255000.],
       [  1604., 242900.],
       [  1962., 259900.],
       [  3890., 573900.],
       [  1100., 249900.],
       [  1458., 464500.],
       [  2526., 469000.],
       [  2200., 475000.],
       [  2637., 299900.],
       [  1839., 349900.],
       [  1000., 169900.],
       [  2040., 314900.],
       [  3137., 579900.],
       [  1811., 285900.],
       [  1437., 249900.],
       [  1239., 229900.],
       [  2132., 345000.],
       [  4215., 549000.],
       [  2162., 287000.],
       [  1664., 368500.],
       [  2238., 329900.],
       [  2567., 314000.],
       [  1200., 299000.],
       [   852., 179900.],
       [  1852., 299900.],
       [  1203., 239500.]])

In [31]:

#切片，同样适用于dnarray，list切片语法为[start : end : step]，应用到多维则是[start : end : step, start : end : step, ……]
dnarr = np.arange(25).reshape(5,5)
#选取头两列、所有行的数据
dnarr[:, :2]

Out[31]:

array([[ 0,  1],
       [ 5,  6],
       [10, 11],
       [15, 16],
       [20, 21]])

In [34]:

#坐标为数组
x = [1, 2, 3, 4]
y = [0, 1, 0, 1]
dnarr[x, y]

Out[34]:

array([ 5, 11, 15, 21])

In [35]:

#广播(Broadcast)是 numpy 对不同形状(shape)的数组进行数值计算的方式，当运算中的 2 个数组的形状不同时，numpy 将自动触发广播机制
a = np.array([[ 0, 0, 0],
           [10,10,10],
           [20,20,20],
           [30,30,30]])
b = np.array([1,2,3])
a + b

Out[35]:

array([[ 1,  2,  3],
       [11, 12, 13],
       [21, 22, 23],
       [31, 32, 33]])

In [40]:

#翻转矩阵
a.T

Out[40]:

array([[ 0, 10, 20, 30],
       [ 0, 10, 20, 30],
       [ 0, 10, 20, 30]])

In [56]:

#数组的运算，对数组中的所有元素进行计算，非常方便
a = np.random.rand(3,3)
b = np.random.rand(3,3)
a + b
a - b
a * b
a / b
a * 2
a + 2

Out[56]:

array([[2.37861674, 2.24048424, 2.06956704],
       [2.21845504, 2.45944021, 2.56521539],
       [2.72539798, 2.51533669, 2.7490997 ]])

In [69]:

#统计函数

#按轴统计最大最小值
np.amax(a, 1)
np.amin(a, 0)

#最大最小差值
np.ptp(a, 0)
#中位数
np.median(a, 1)
#平均值
np.mean(a, 1)

#方差
np.var(a, 1)
#标准差
np.std(a, 1)

Out[69]:

array([0.12640543, 0.14510723, 0.10505695])

In [71]:

#矩阵
import numpy.matlib 

numpy.matlib.rand(3,3)

Out[71]:

matrix([[0.51150901, 0.0553782 , 0.99089512],
        [0.51719364, 0.00331011, 0.08076412],
        [0.66316065, 0.94387314, 0.30780752]])

In [ ]: