3-1 jupyter notebook基础
Notbook 示例
Notbook 源码
1 [1]
2 for x in range(5):
3 print('hello world')
4 hello world
5 hello world
6 hello world
7 hello world
8 hello world
9
10 欢迎来到机器学习
11 阿济格的回复考核
12
13 [2]
14 data = [ 2*i for i in range(100) ]
15 print(data)
16 [0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, 72, 74, 76, 78, 80, 82, 84, 86, 88, 90, 92, 94, 96, 98, 100, 102, 104, 106, 108, 110, 112, 114, 116, 118, 120, 122, 124, 126, 128, 130, 132, 134, 136, 138, 140, 142, 144, 146, 148, 150, 152, 154, 156, 158, 160, 162, 164, 166, 168, 170, 172, 174, 176, 178, 180, 182, 184, 186, 188, 190, 192, 194, 196, 198]
17
18 [3]
19 data[-1]
20 198
21 变量一旦在jupyer notebook 中创建即被保存
22
23 [4]
24 len(data)
25 100
26 [8]
27 data[:10]
28 [0, 2, 4, 6, 8, 10, 12, 14, 16, 18]
29 [6]
30 print(5464654)
31 5464654
3-2 jupyter notebook中的魔法命令
Notbook 示例
Notbook 源码
1 [1]
2 %run myscript/hello.py
3 hello machine learning !
4
5 [2]
6 hello('what man')
7 hello what man !
8
9 [3]
10 import mymodule.firstml
11 8848
12 能不能运行
13
14 直接predict(8)不能运行
15
16 通过 import 直接运行的文件，其文件中的函数不能为下节点直接调用
17
18 [4]
19 mymodule.firstml.predict(8)
20 8
21
22 对于上面 mymodule必不可少
23
24 [5]
25 %run mymodule/firstml.py
26 8848
27 能不能运行
28
29 [6]
30 import myscript.hello
31 hello machine learning !
32
33 [7]
34 from mymodule import firstml
35 [8]
36 firstml.predict(9)
37 9
38
39 [9]
40 %timeit L = [ i**2 for i in range(1000) ]
41 497 µs ± 27.6 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)
42
43 [10]
44 %timeit L = [ i**2 for i in range(1000000) ]
45 638 ms ± 96.8 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
46
47 [11]
48 %timeit L = [ i**2 for i in range(10) ]
49 5.21 µs ± 255 ns per loop (mean ± std. dev. of 7 runs, 100,000 loops each)
50
51 [12]
52 %%timeit
53 L = [ ]
54 for n in range(1000):
55 L.append( n ** 2)
56 566 µs ± 28.5 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)
57
58 [13]
59 %time L = [ i**2 for i in range(1000) ]
60 CPU times: total: 0 ns
61 Wall time: 0 ns
62
63 [14]
64 %%time
65 L = [ ]
66 for n in range(1000):
67 L.append( n ** 2)
68 CPU times: total: 0 ns
69 Wall time: 0 ns
70
71 [15]
72 import random
73 L = [random.random() for n in range(100000)]
74 %timeit L.sort()
75 2.48 ms ± 223 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)
76
77 random.random()方法返回一个随机数，其在0至1的范围之内，以下是其具体用法： import random print ("随机数: ", random.random())
78
79 输出结果：0.22867521257116
80
81 sort() 函数是序列的内部函数，用于对原列表进行排序，如果指定参数，则使用比较函数指定的比较函数。
82
83 把L原地排序，也就是使用后并不是返回一个有序的序列副本，而是把当前序列变得有序。
84
85 2、基本形式
86
87 l.sort()（l是一种列表）
88
89 [16]
90 L = [random.random() for n in range(100000)]
91 %time L.sort()
92 CPU times: total: 31.2 ms
93 Wall time: 32.9 ms
94
95 [17]
96 %time L.sort()
97 CPU times: total: 0 ns
98 Wall time: 2 ms
99
100 [18]
101 %lsmagic
102 Available line magics:
103 %alias %alias_magic %autoawait %autocall %automagic %autosave %bookmark %cd %clear %cls %colors %conda %config %connect_info %copy %ddir %debug %dhist %dirs %doctest_mode %echo %ed %edit %env %gui %hist %history %killbgscripts %ldir %less %load %load_ext %loadpy %logoff %logon %logstart %logstate %logstop %ls %lsmagic %macro %magic %matplotlib %mkdir %more %notebook %page %pastebin %pdb %pdef %pdoc %pfile %pinfo %pinfo2 %pip %popd %pprint %precision %prun %psearch %psource %pushd %pwd %pycat %pylab %qtconsole %quickref %recall %rehashx %reload_ext %ren %rep %rerun %reset %reset_selective %rmdir %run %save %sc %set_env %store %sx %system %tb %time %timeit %unalias %unload_ext %who %who_ls %whos %xdel %xmode
104
105 Available cell magics:
106 %%! %%HTML %%SVG %%bash %%capture %%cmd %%debug %%file %%html %%javascript %%js %%latex %%markdown %%perl %%prun %%pypy %%python %%python2 %%python3 %%ruby %%script %%sh %%svg %%sx %%system %%time %%timeit %%writefile
107
108 Automagic is ON, % prefix IS NOT needed for line magics.
109 [19]
110 %run?
3-3 Numpy数据基础and3-4 创建numpy数组和矩阵
Notbook 示例
Notbook 源码
1 [1]
2 import numpy
3 [2]
4 numpy.__version__
5 '1.21.5'
6 [3]
7 import numpy as np
8 [4]
9 np.__version__
10 '1.21.5'
11 python list 的特点
12
13 [5]
14 L = [ i for i in range(10)]
15 L
16 [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
17 [6]
18 L[5]
19 5
20 [7]
21 L[5] = 100
22 L
23 [0, 1, 2, 3, 4, 100, 6, 7, 8, 9]
24 [8]
25 L[5] = 'machine learning'
26 L
27 [0, 1, 2, 3, 4, 'machine learning', 6, 7, 8, 9]
28 上面单双引号均可
29
30 [9]
31 import array
32 [10]
33 arr = array.array('i',[ i for i in range(10)])
34 arr
35 array('i', [0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
36 [11]
37 arr[5]
38 5
39 [12]
40 arr[5] = 100
41 arr[5]
42 100
43 arr[5] = 'machine learing '报错 array.array，限定数据类型。限制了灵活性，相对速度比较高；同时array只是将存储的数据看成数组或二维数组，而数组并没有看成矩阵，也没有配备向量或矩阵相关的运算； myArray = array.array('i', [i for i range(10)])
44
45 numpy.array
46 [13]
47 nparr = np.array([i for i in range(10)])
48 nparr
49 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
50 [14]
51 nparr[5]
52 5
53 [15]
54 nparr[5] = 10
55 nparr
56 array([ 0, 1, 2, 3, 4, 10, 6, 7, 8, 9])
57 nparry[5] = "machine learning" 将报错
58
59 [16]
60 nparr.dtype
61 dtype('int32')
62 [17]
63 nparr[5] = 5.0
64 [18]
65 nparr
66 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
67 [19]
68 nparr.dtype
69 dtype('int32')
70 [20]
71 nparr[5] = 3.14
72 nparr
73 array([0, 1, 2, 3, 4, 3, 6, 7, 8, 9])
74 [21]
75 nparr2 = np.array([1, 2, 3.0])
76 nparr2.dtype
77 dtype('float64')
78 [22]
79 np.zeros(10)
80 array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0.])
81 [23]
82 np.zeros(10).dtype
83 dtype('float64')
84 [24]
85 np.zeros(10,dtype = int)
86 array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0])
87 [25]
88 np.zeros((3,5))
89 array([[0., 0., 0., 0., 0.],
90 [0., 0., 0., 0., 0.],
91 [0., 0., 0., 0., 0.]])
92 [58]
93 np.zeros(shape = (3,5),dtype = int)
94 array([[0, 0, 0, 0, 0],
95 [0, 0, 0, 0, 0],
96 [0, 0, 0, 0, 0]])
97 [27]
98 np.ones(10)
99 array([1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])
100 [28]
101 np.ones((3,5))
102 array([[1., 1., 1., 1., 1.],
103 [1., 1., 1., 1., 1.],
104 [1., 1., 1., 1., 1.]])
105 [29]
106 np.full((3,6),666)
107 array([[666, 666, 666, 666, 666, 666],
108 [666, 666, 666, 666, 666, 666],
109 [666, 666, 666, 666, 666, 666]])
110 [30]
111 np.full(shape = (3,6),fill_value = 666)
112 array([[666, 666, 666, 666, 666, 666],
113 [666, 666, 666, 666, 666, 666],
114 [666, 666, 666, 666, 666, 666]])
115 fill_value 只有一个下划线
116
117 [31]
118 np.full(fill_value = 666,shape = (3,6))
119 array([[666, 666, 666, 666, 666, 666],
120 [666, 666, 666, 666, 666, 666],
121 [666, 666, 666, 666, 666, 666]])
122 [32]
123 np.full(fill_value = 6,shape = (10))
124 array([6, 6, 6, 6, 6, 6, 6, 6, 6, 6])
125 [33]
126 [i for i in range(0,20,2)]
127 [0, 2, 4, 6, 8, 10, 12, 14, 16, 18]
128 [34]
129 np.arange(0,20,2)
130 array([ 0, 2, 4, 6, 8, 10, 12, 14, 16, 18])
131 [35]
132 np.arange(0,1,0.2)
133 array([0. , 0.2, 0.4, 0.6, 0.8])
134 在python中range步长只能为整数
135
136 [36]
137 np.arange(0,10)
138 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
139 [37]
140 np.arange(10)
141 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
142 linspace
143 [38]
144 np.linspace(0,20,10)
145 array([ 0. , 2.22222222, 4.44444444, 6.66666667, 8.88888889,
146 11.11111111, 13.33333333, 15.55555556, 17.77777778, 20. ])
147 [39]
148 np.linspace(0,20,11)
149 array([ 0., 2., 4., 6., 8., 10., 12., 14., 16., 18., 20.])
150 random
151 [40]
152 np.random.randint(0,10)
153 6
154 [41]
155 np.random.randint(0,10,10)
156 array([7, 0, 8, 7, 7, 4, 1, 4, 4, 8])
157 前闭后开，对于上面永远娶不到10
158
159 [42]
160 np.random.randint(0,1,10)
161 array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0])
162 [43]
163 np.random.randint(4,8,size = 10)
164 array([7, 6, 6, 7, 5, 6, 4, 6, 5, 7])
165 [44]
166 np.random.randint(4,8,size = (3,5))
167 array([[4, 4, 5, 7, 6],
168 [7, 7, 4, 7, 7],
169 [6, 5, 5, 6, 7]])
170 [45]
171 np.random.randint(4,8,size = (3,5))
172 array([[7, 4, 6, 6, 6],
173 [7, 4, 5, 6, 5],
174 [4, 4, 6, 6, 7]])
175 [46]
176 np.random.seed(666)
177 [47]
178 np.random.randint(4,8,size = (3,5))
179 array([[4, 6, 5, 6, 6],
180 [6, 5, 6, 4, 5],
181 [7, 6, 7, 4, 7]])
182 [48]
183 np.random.seed(666)
184 np.random.randint(4,8,size = (3,5))
185 array([[4, 6, 5, 6, 6],
186 [6, 5, 6, 4, 5],
187 [7, 6, 7, 4, 7]])
188 [49]
189 np.random.random()
190 0.2811684913927954
191 最后一个random必须加括号
192
193 [50]
194 np.random.random(10)
195 array([0.46284169, 0.23340091, 0.76706421, 0.81995656, 0.39747625,
196 0.31644109, 0.15551206, 0.73460987, 0.73159555, 0.8578588 ])
197 [51]
198 np.random.random((3,5))
199 array([[0.76741234, 0.95323137, 0.29097383, 0.84778197, 0.3497619 ],
200 [0.92389692, 0.29489453, 0.52438061, 0.94253896, 0.07473949],
201 [0.27646251, 0.4675855 , 0.31581532, 0.39016259, 0.26832981]])
202 [52]
203 np.random.normal()
204 0.7760516793129695
205 [53]
206 np.random.normal(10,100)
207 128.06359754812632
208 均值为10，方差为100
209
210 [54]
211 np.random.normal(0,1,(3,5))
212 array([[ 0.06102404, 1.07856138, -0.79783572, 1.1701326 , 0.1121217 ],
213 [ 0.03185388, -0.19206285, 0.78611284, -1.69046314, -0.98873907],
214 [ 0.31398563, 0.39638567, 0.57656584, -0.07019407, 0.91250436]])
215 [59]
216 np.random.normal?
217 [60]
218 np.random?
219 [57]
220 help(np.random.normal)
221 Help on built-in function normal:
222
223 normal(...) method of numpy.random.mtrand.RandomState instance
224 normal(loc=0.0, scale=1.0, size=None)
225
226 Draw random samples from a normal (Gaussian) distribution.
227
228 The probability density function of the normal distribution, first
229 derived by De Moivre and 200 years later by both Gauss and Laplace
230 independently [2]_, is often called the bell curve because of
231 its characteristic shape (see the example below).
232
233 The normal distributions occurs often in nature. For example, it
234 describes the commonly occurring distribution of samples influenced
235 by a large number of tiny, random disturbances, each with its own
236 unique distribution [2]_.
237
238 .. note::
239 New code should use the ``normal`` method of a ``default_rng()``
240 instance instead; please see the :ref:`random-quick-start`.
241
242 Parameters
243 ----------
244 loc : float or array_like of floats
245 Mean ("centre") of the distribution.
246 scale : float or array_like of floats
247 Standard deviation (spread or "width") of the distribution. Must be
248 non-negative.
249 size : int or tuple of ints, optional
250 Output shape. If the given shape is, e.g., ``(m, n, k)``, then
251 ``m * n * k`` samples are drawn. If size is ``None`` (default),
252 a single value is returned if ``loc`` and ``scale`` are both scalars.
253 Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn.
254
255 Returns
256 -------
257 out : ndarray or scalar
258 Drawn samples from the parameterized normal distribution.
259
260 See Also
261 --------
262 scipy.stats.norm : probability density function, distribution or
263 cumulative density function, etc.
264 Generator.normal: which should be used for new code.
265
266 Notes
267 -----
268 The probability density for the Gaussian distribution is
269
270 .. math:: p(x) = \frac{1}{\sqrt{ 2 \pi \sigma^2 }}
271 e^{ - \frac{ (x - \mu)^2 } {2 \sigma^2} },
272
273 where :math:`\mu` is the mean and :math:`\sigma` the standard
274 deviation. The square of the standard deviation, :math:`\sigma^2`,
275 is called the variance.
276
277 The function has its peak at the mean, and its "spread" increases with
278 the standard deviation (the function reaches 0.607 times its maximum at
279 :math:`x + \sigma` and :math:`x - \sigma` [2]_). This implies that
280 normal is more likely to return samples lying close to the mean, rather
281 than those far away.
282
283 References
284 ----------
285 .. [1] Wikipedia, "Normal distribution",
286 https://en.wikipedia.org/wiki/Normal_distribution
287 .. [2] P. R. Peebles Jr., "Central Limit Theorem" in "Probability,
288 Random Variables and Random Signal Principles", 4th ed., 2001,
289 pp. 51, 51, 125.
290
291 Examples
292 --------
293 Draw samples from the distribution:
294
295 >>> mu, sigma = 0, 0.1 # mean and standard deviation
296 >>> s = np.random.normal(mu, sigma, 1000)
297
298 Verify the mean and the variance:
299
300 >>> abs(mu - np.mean(s))
301 0.0 # may vary
302
303 >>> abs(sigma - np.std(s, ddof=1))
304 0.1 # may vary
305
306 Display the histogram of the samples, along with
307 the probability density function:
308
309 >>> import matplotlib.pyplot as plt
310 >>> count, bins, ignored = plt.hist(s, 30, density=True)
311 >>> plt.plot(bins, 1/(sigma * np.sqrt(2 * np.pi)) *
312 ... np.exp( - (bins - mu)**2 / (2 * sigma**2) ),
313 ... linewidth=2, color='r')
314 >>> plt.show()
315
316 Two-by-four array of samples from N(3, 6.25):
317
318 >>> np.random.normal(3, 2.5, size=(2, 4))
319 array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], # random
320 [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) # random
3-5 Numpy数组的基本操作and3-6 Numpy数组的合并与分割
Notbook 示例
Notbook 源码
1 [1] 2 import numpy as np 3 [2] 4 x = np.arange(10) 5 x 6 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 7 [3] 8 X = np.arange(15).reshape(3,5) 9 X 10 array([[ 0, 1, 2, 3, 4], 11 [ 5, 6, 7, 8, 9], 12 [10, 11, 12, 13, 14]]) 13 基本属性 14 [4] 15 x.ndim 16 1 17 [5] 18 X.ndim 19 2 20 [6] 21 x.shape 22 (10,) 23 [7] 24 X.shape 25 (3, 5) 26 [8] 27 x.size 28 10 29 [9] 30 X.size 31 15 32 Numpy.array的数据访问 33 [10] 34 x 35 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 36 [11] 37 x[0] 38 0 39 [12] 40 x[-1] 41 9 42 [13] 43 X 44 array([[ 0, 1, 2, 3, 4], 45 [ 5, 6, 7, 8, 9], 46 [10, 11, 12, 13, 14]]) 47 [14] 48 X[0][0] 49 0 50 [15] 51 X[(2,2)] 52 12 53 [16] 54 X[2,2] 55 12 56 [17] 57 X[2][2] 58 12 59 [18] 60 X[:2][:2] 61 array([[0, 1, 2, 3, 4], 62 [5, 6, 7, 8, 9]]) 63 注意与切片方式区分 64 65 [19] 66 x[0:5] 67 array([0, 1, 2, 3, 4]) 68 [20] 69 x[:5] 70 array([0, 1, 2, 3, 4]) 71 [21] 72 x[5:] 73 array([5, 6, 7, 8, 9]) 74 [22] 75 x[::2] 76 array([0, 2, 4, 6, 8]) 77 [23] 78 x[::-1] 79 array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0]) 80 [24] 81 X 82 array([[ 0, 1, 2, 3, 4], 83 [ 5, 6, 7, 8, 9], 84 [10, 11, 12, 13, 14]]) 85 [25] 86 X[:2,:3] 87 array([[0, 1, 2], 88 [5, 6, 7]]) 89 [26] 90 x 91 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 92 [27] 93 x[:2][:3] 94 array([0, 1]) 95 [28] 96 X[:2][:3] 97 array([[0, 1, 2, 3, 4], 98 [5, 6, 7, 8, 9]]) 99 [29] 100 X[:2] 101 array([[0, 1, 2, 3, 4], 102 [5, 6, 7, 8, 9]]) 103 [30] 104 X[:2][:3] 105 array([[0, 1, 2, 3, 4], 106 [5, 6, 7, 8, 9]]) 107 X[:2][:3]在X[:2]的基础上取值，还是按行取，欲取三个只有两个，故全部取出 108 109 [31] 110 X 111 array([[ 0, 1, 2, 3, 4], 112 [ 5, 6, 7, 8, 9], 113 [10, 11, 12, 13, 14]]) 114 [32] 115 X[:2,::2] #::2 116 array([[0, 2, 4], 117 [5, 7, 9]]) 118 [33] 119 X[::-1,::-1] 120 array([[14, 13, 12, 11, 10], 121 [ 9, 8, 7, 6, 5], 122 [ 4, 3, 2, 1, 0]]) 123 [34] 124 X[0] 125 array([0, 1, 2, 3, 4]) 126 [35] 127 X[0,::] #对于此行及以下三行，：与：：效果相同 128 array([0, 1, 2, 3, 4]) 129 [36] 130 X[::,0] 131 array([ 0, 5, 10]) 132 [37] 133 X[0,::].ndim 134 1 135 [38] 136 X[::,0] 137 array([ 0, 5, 10]) 138 [39] 139 X[:,0].ndim 140 1 141 [40] 142 subx = X[:2,:3] 143 subx 144 array([[0, 1, 2], 145 [5, 6, 7]]) 146 [41] 147 subx[0][0]=100 148 subx 149 array([[100, 1, 2], 150 [ 5, 6, 7]]) 151 [42] 152 X 153 array([[100, 1, 2, 3, 4], 154 [ 5, 6, 7, 8, 9], 155 [ 10, 11, 12, 13, 14]]) 156 相对于python 创建一个子矩阵，numpy则直接引用 157 158 [43] 159 X[0][0] = 0 160 X 161 array([[ 0, 1, 2, 3, 4], 162 [ 5, 6, 7, 8, 9], 163 [10, 11, 12, 13, 14]]) 164 [44] 165 subx 166 array([[0, 1, 2], 167 [5, 6, 7]]) 168 [45] 169 subx = X[:2,:3].copy() 170 subx 171 array([[0, 1, 2], 172 [5, 6, 7]]) 173 [46] 174 subx[0][0] = 100 175 subx 176 array([[100, 1, 2], 177 [ 5, 6, 7]]) 178 [47] 179 X 180 array([[ 0, 1, 2, 3, 4], 181 [ 5, 6, 7, 8, 9], 182 [10, 11, 12, 13, 14]]) 183 reshape 184 [48] 185 x.shape 186 (10,) 187 [49] 188 x.ndim 189 1 190 [50] 191 x.reshape(2,5) 192 array([[0, 1, 2, 3, 4], 193 [5, 6, 7, 8, 9]]) 194 [51] 195 x 196 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 197 [52] 198 A = x.reshape(2,5) 199 A 200 array([[0, 1, 2, 3, 4], 201 [5, 6, 7, 8, 9]]) 202 [53] 203 x 204 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 205 [54] 206 B = x.reshape(1,10) 207 B 208 array([[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]]) 209 [55] 210 B.ndim 211 2 212 [56] 213 B.shape 214 (1, 10) 215 [57] 216 x.shape 217 (10,) 218 [58] 219 x.reshape(10,-1) 220 array([[0], 221 [1], 222 [2], 223 [3], 224 [4], 225 [5], 226 [6], 227 [7], 228 [8], 229 [9]]) 230 [59] 231 x.reshape(-1,10) 232 array([[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]]) 233 [60] 234 x.reshape(2,-1) 235 array([[0, 1, 2, 3, 4], 236 [5, 6, 7, 8, 9]]) 237 合并操作 238 [61] 239 x = np.array([1,2,3]) 240 y = np.array([3,2,1]) 241 [62] 242 x 243 array([1, 2, 3]) 244 [63] 245 y 246 array([3, 2, 1]) 247 [64] 248 np.concatenate([x,y]) 249 array([1, 2, 3, 3, 2, 1]) 250 [65] 251 z = np.array([666,666,666]) 252 [66] 253 np.concatenate([x,y,z]) 254 array([ 1, 2, 3, 3, 2, 1, 666, 666, 666]) 255 [67] 256 A = np.array([[1,2,3],[4,5,6]]) 257 A 258 array([[1, 2, 3], 259 [4, 5, 6]]) 260 [68] 261 np.concatenate([A,A]) 262 array([[1, 2, 3], 263 [4, 5, 6], 264 [1, 2, 3], 265 [4, 5, 6]]) 266 [69] 267 np.concatenate([A,A],axis = 1) 268 array([[1, 2, 3, 1, 2, 3], 269 [4, 5, 6, 4, 5, 6]]) 270 [70] 271 np.concatenate([A,A],axis = 0) 272 array([[1, 2, 3], 273 [4, 5, 6], 274 [1, 2, 3], 275 [4, 5, 6]]) 276 [71] 277 z 278 array([666, 666, 666]) 279 [72] 280 A 281 array([[1, 2, 3], 282 [4, 5, 6]]) 283 [73] 284 # np.vstack([A,z]) vstack 为操作向量提供方便 285 np.concatenate([A,z],axis = 0) 将报错，因为维度不一样 286 287 [74] 288 np.concatenate([ A, z.reshape(1,-1)]) 289 array([[ 1, 2, 3], 290 [ 4, 5, 6], 291 [666, 666, 666]]) 292 这里的z只是一维的向量，不能直接和A进行concatenate操作， 需要先转换为1 * 3 的矩阵 293 294 [75] 295 A 296 array([[1, 2, 3], 297 [4, 5, 6]]) 298 [76] 299 A2 = np.concatenate([A,z.reshape([1,-1])]) 300 A2 301 array([[ 1, 2, 3], 302 [ 4, 5, 6], 303 [666, 666, 666]]) 304 [77] 305 np.vstack([A,z]) 306 array([[ 1, 2, 3], 307 [ 4, 5, 6], 308 [666, 666, 666]]) 309 [78] 310 B = np.full((2,2),100) 311 B 312 array([[100, 100], 313 [100, 100]]) 314 [79] 315 np.hstack([A,B]) 316 array([[ 1, 2, 3, 100, 100], 317 [ 4, 5, 6, 100, 100]]) 318 分割操作 319 [80] 320 x = np.arange(10) 321 x 322 array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 323 [81] 324 x1,x2,x3 = np.split(x,[3,7]) 325 [82] 326 x1 327 array([0, 1, 2]) 328 [83] 329 x2 330 array([3, 4, 5, 6]) 331 [84] 332 x3 333 array([7, 8, 9]) 334 [85] 335 x1,x2 = np.split(x,[5]) 336 [86] 337 x1 338 array([0, 1, 2, 3, 4]) 339 [87] 340 x2 341 array([5, 6, 7, 8, 9]) 342 [116] 343 A = np.arange(16).reshape(4,4) 344 A 345 array([[ 0, 1, 2, 3], 346 [ 4, 5, 6, 7], 347 [ 8, 9, 10, 11], 348 [12, 13, 14, 15]]) 349 [115] 350 A.shape 351 (2, 8) 352 [89] 353 A1,A2 = np.split(A,[2]) 354 [90] 355 A1 356 array([[0, 1, 2, 3], 357 [4, 5, 6, 7]]) 358 如果A1 = np.split(A,[2])，少一个A2参数则 [array([[0, 1, 2, 3], [4, 5, 6, 7]]), array([[ 8, 9, 10, 11], [12, 13, 14, 15]])] 359 360 [91] 361 A2 362 array([[ 8, 9, 10, 11], 363 [12, 13, 14, 15]]) 364 [92] 365 A1,A2 = np.split(A,[2],axis = 1) 366 [93] 367 A1 368 array([[ 0, 1], 369 [ 4, 5], 370 [ 8, 9], 371 [12, 13]]) 372 [94] 373 A2 374 array([[ 2, 3], 375 [ 6, 7], 376 [10, 11], 377 [14, 15]]) 378 [95] 379 A 380 array([[ 0, 1, 2, 3], 381 [ 4, 5, 6, 7], 382 [ 8, 9, 10, 11], 383 [12, 13, 14, 15]]) 384 [96] 385 upper,lower = np.vsplit(A,[2]) 386 [97] 387 upper 388 array([[0, 1, 2, 3], 389 [4, 5, 6, 7]]) 390 [98] 391 lower 392 array([[ 8, 9, 10, 11], 393 [12, 13, 14, 15]]) 394 [99] 395 left,right = np.hsplit(A,[2]) 396 [100] 397 left 398 array([[ 0, 1], 399 [ 4, 5], 400 [ 8, 9], 401 [12, 13]]) 402 [101] 403 right 404 array([[ 2, 3], 405 [ 6, 7], 406 [10, 11], 407 [14, 15]]) 408 [102] 409 data = np.arange(16).reshape(4,4) 410 data 411 array([[ 0, 1, 2, 3], 412 [ 4, 5, 6, 7], 413 [ 8, 9, 10, 11], 414 [12, 13, 14, 15]]) 415 [103] 416 x,y = np.hsplit(data,[-1]) 417 [104] 418 x 419 array([[ 0, 1, 2], 420 [ 4, 5, 6], 421 [ 8, 9, 10], 422 [12, 13, 14]]) 423 [105] 424 y 425 array([[ 3], 426 [ 7], 427 [11], 428 [15]]) 429 [106] 430 y[:,0] 431 array([ 3, 7, 11, 15])
3-7 Numpy中的矩阵运算
Notbook 示例
Notbook 源码
1 [1] 2 n = 10 3 L = [ i for i in range(n)] 4 [2] 5 2 * L 6 [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9] 7 [3] 8 A = [] 9 for e in L: 10 A.append(2*e) 11 A 12 [0, 2, 4, 6, 8, 10, 12, 14, 16, 18] 13 [4] 14 n = 1000000 15 L = [ i for i in range(n)] 16 [5] 17 %%time 18 A = [] 19 for e in L: 20 A.append(2*e) 21 CPU times: total: 375 ms 22 Wall time: 449 ms 23 24 [6] 25 %%time 26 A = [e*2 for e in L] 27 CPU times: total: 172 ms 28 Wall time: 209 ms 29 30 [7] 31 import numpy as np 32 L = np.arange(n) 33 %%time A = [e*2 for e in L] 把numpy L = np.arange(n) CPU times: total: 453 ms Wall time: 475 ms %%time百分号与time之间不能有空格 34 35 [8] 36 %%time 37 A = np.array(2*e for e in L) 38 CPU times: total: 31.2 ms 39 Wall time: 22.9 ms 40 41 %%time百分号与time之间不能有空格 42 43 [9] 44 %time 45 A = 2*L 46 CPU times: total: 0 ns 47 Wall time: 0 ns 48 49 [10] 50 A 51 array([ 0, 2, 4, ..., 1999994, 1999996, 1999998]) 52 [11] 53 n = 10 54 L = np.arange(n) 55 2*L 56 array([ 0, 2, 4, 6, 8, 10, 12, 14, 16, 18]) 57 universal function 58 [12] 59 x = np.arange(1,16).reshape(3,5) 60 x 61 array([[ 1, 2, 3, 4, 5], 62 [ 6, 7, 8, 9, 10], 63 [11, 12, 13, 14, 15]]) 64 [13] 65 x+1 66 array([[ 2, 3, 4, 5, 6], 67 [ 7, 8, 9, 10, 11], 68 [12, 13, 14, 15, 16]]) 69 [14] 70 x-1 71 array([[ 0, 1, 2, 3, 4], 72 [ 5, 6, 7, 8, 9], 73 [10, 11, 12, 13, 14]]) 74 [15] 75 x*2 76 array([[ 2, 4, 6, 8, 10], 77 [12, 14, 16, 18, 20], 78 [22, 24, 26, 28, 30]]) 79 [16] 80 x/2 81 array([[0.5, 1. , 1.5, 2. , 2.5], 82 [3. , 3.5, 4. , 4.5, 5. ], 83 [5.5, 6. , 6.5, 7. , 7.5]]) 84 [17] 85 x//2 86 array([[0, 1, 1, 2, 2], 87 [3, 3, 4, 4, 5], 88 [5, 6, 6, 7, 7]], dtype=int32) 89 [18] 90 x ** 2 91 array([[ 1, 4, 9, 16, 25], 92 [ 36, 49, 64, 81, 100], 93 [121, 144, 169, 196, 225]], dtype=int32) 94 [19] 95 x % 2 96 array([[1, 0, 1, 0, 1], 97 [0, 1, 0, 1, 0], 98 [1, 0, 1, 0, 1]], dtype=int32) 99 [20] 100 1 / x 101 array([[1. , 0.5 , 0.33333333, 0.25 , 0.2 ], 102 [0.16666667, 0.14285714, 0.125 , 0.11111111, 0.1 ], 103 [0.09090909, 0.08333333, 0.07692308, 0.07142857, 0.06666667]]) 104 [21] 105 np.abs(x) 106 array([[ 1, 2, 3, 4, 5], 107 [ 6, 7, 8, 9, 10], 108 [11, 12, 13, 14, 15]]) 109 [22] 110 np.sin(x) 111 array([[ 0.84147098, 0.90929743, 0.14112001, -0.7568025 , -0.95892427], 112 [-0.2794155 , 0.6569866 , 0.98935825, 0.41211849, -0.54402111], 113 [-0.99999021, -0.53657292, 0.42016704, 0.99060736, 0.65028784]]) 114 [23] 115 np.cos(x) 116 array([[ 0.54030231, -0.41614684, -0.9899925 , -0.65364362, 0.28366219], 117 [ 0.96017029, 0.75390225, -0.14550003, -0.91113026, -0.83907153], 118 [ 0.0044257 , 0.84385396, 0.90744678, 0.13673722, -0.75968791]]) 119 [24] 120 np.tan(x) 121 array([[ 1.55740772e+00, -2.18503986e+00, -1.42546543e-01, 122 1.15782128e+00, -3.38051501e+00], 123 [-2.91006191e-01, 8.71447983e-01, -6.79971146e+00, 124 -4.52315659e-01, 6.48360827e-01], 125 [-2.25950846e+02, -6.35859929e-01, 4.63021133e-01, 126 7.24460662e+00, -8.55993401e-01]]) 127 [25] 128 np.exp(x) 129 array([[2.71828183e+00, 7.38905610e+00, 2.00855369e+01, 5.45981500e+01, 130 1.48413159e+02], 131 [4.03428793e+02, 1.09663316e+03, 2.98095799e+03, 8.10308393e+03, 132 2.20264658e+04], 133 [5.98741417e+04, 1.62754791e+05, 4.42413392e+05, 1.20260428e+06, 134 3.26901737e+06]]) 135 e的x次方 136 137 [26] 138 np.power(3,x) 139 array([[ 3, 9, 27, 81, 243], 140 [ 729, 2187, 6561, 19683, 59049], 141 [ 177147, 531441, 1594323, 4782969, 14348907]], dtype=int32) 142 [27] 143 3 ** x 144 array([[ 3, 9, 27, 81, 243], 145 [ 729, 2187, 6561, 19683, 59049], 146 [ 177147, 531441, 1594323, 4782969, 14348907]], dtype=int32) 147 [28] 148 np.log(x) 149 array([[0. , 0.69314718, 1.09861229, 1.38629436, 1.60943791], 150 [1.79175947, 1.94591015, 2.07944154, 2.19722458, 2.30258509], 151 [2.39789527, 2.48490665, 2.56494936, 2.63905733, 2.7080502 ]]) 152 直接调用log相对于以e为底 153 154 [29] 155 np.log2(x) 156 array([[0. , 1. , 1.5849625 , 2. , 2.32192809], 157 [2.5849625 , 2.80735492, 3. , 3.169925 , 3.32192809], 158 [3.45943162, 3.5849625 , 3.70043972, 3.80735492, 3.9068906 ]]) 159 [30] 160 np.log10(x) 161 array([[0. , 0.30103 , 0.47712125, 0.60205999, 0.69897 ], 162 [0.77815125, 0.84509804, 0.90308999, 0.95424251, 1. ], 163 [1.04139269, 1.07918125, 1.11394335, 1.14612804, 1.17609126]]) 164 矩阵运算 165 [31] 166 A = np.arange(4).reshape(2,2) 167 A 168 array([[0, 1], 169 [2, 3]]) 170 如果 A = np.arange(2,2) A 则有 array([], dtype=int32) 171 172 [32] 173 B = np.full((2,2),10) 174 B 175 array([[10, 10], 176 [10, 10]]) 177 [33] 178 A + B 179 array([[10, 11], 180 [12, 13]]) 181 [34] 182 A - B 183 array([[-10, -9], 184 [ -8, -7]]) 185 [35] 186 A * B 187 array([[ 0, 10], 188 [20, 30]]) 189 [36] 190 A / B 191 array([[0. , 0.1], 192 [0.2, 0.3]]) 193 [37] 194 A.dot(B) 195 array([[10, 10], 196 [50, 50]]) 197 [38] 198 A 199 array([[0, 1], 200 [2, 3]]) 201 [39] 202 A.T 203 array([[0, 2], 204 [1, 3]]) 205 C = np.full((3,3),666) A + C A.dot(C) 对于上代码后两个等式将不能运行,需要符合矩阵运算规则 206 207 [40] 208 V = np.array([1,2]) 209 V 210 array([1, 2]) 211 V = np.array((1,2)) V 运行结果相同 212 213 [41] 214 A 215 array([[0, 1], 216 [2, 3]]) 217 [42] 218 V + A 219 array([[1, 3], 220 [3, 5]]) 221 若A = np.arange(6).reshape(3,2) A array([[0, 1], [2, 3], [4, 5]]) 则np.vstack([v] * A.shape[0]) array([[1, 2], [1, 2], [1, 2]]) 222 223 [43] 224 np.vstack([V] * A.shape[0]) 225 array([[1, 2], 226 [1, 2]]) 227 A.shape[0]表示以A的基本单元为基础的高度 228 229 [44] 230 np.vstack([V] * A.shape[0])+A 231 array([[1, 3], 232 [3, 5]]) 233 [45] 234 V 235 array([1, 2]) 236 [46] 237 np.tile(V,(2,1)) + A # 横向重复两次，纵向重复一次 238 array([[1, 3], 239 [3, 5]]) 240 [47] 241 V 242 array([1, 2]) 243 [48] 244 A 245 array([[0, 1], 246 [2, 3]]) 247 [49] 248 V * A 249 array([[0, 2], 250 [2, 6]]) 251 [50] 252 V.dot(A) 253 array([4, 7]) 254 [51] 255 A.dot(V) 256 array([2, 8]) 257 对于一个矩阵和一个向量的乘法，numpy会自动将向量转成合适的形式 258 259 矩阵的逆 260 [52] 261 np.linalg.inv(A) 262 array([[-1.5, 0.5], 263 [ 1. , 0. ]]) 264 [53] 265 invA = np.linalg.inv(A) 266 [54] 267 A.dot(invA) 268 array([[1., 0.], 269 [0., 1.]]) 270 [55] 271 invA.dot(A) 272 array([[1., 0.], 273 [0., 1.]]) 274 [56] 275 X = np.arange(16).reshape(2,8) 276 X 277 array([[ 0, 1, 2, 3, 4, 5, 6, 7], 278 [ 8, 9, 10, 11, 12, 13, 14, 15]]) 279 np.linalg.inv(A)将报错 280 281 [57] 282 pinvX = np.linalg.pinv(X) 283 [58] 284 pinvX 285 array([[-1.35416667e-01, 5.20833333e-02], 286 [-1.01190476e-01, 4.16666667e-02], 287 [-6.69642857e-02, 3.12500000e-02], 288 [-3.27380952e-02, 2.08333333e-02], 289 [ 1.48809524e-03, 1.04166667e-02], 290 [ 3.57142857e-02, -3.46944695e-18], 291 [ 6.99404762e-02, -1.04166667e-02], 292 [ 1.04166667e-01, -2.08333333e-02]]) 293 [59] 294 pinvX.shape 295 (8, 2) 296 [60] 297 X.dot(pinvX) 298 array([[ 1.00000000e+00, -1.94289029e-16], 299 [ 6.66133815e-16, 1.00000000e+00]])
3-8 Numpy中的聚合运算and3-9 Numpy中的arg运算
Notbook 示例
Notbook 源码
1 [1] 2 import numpy as np 3 L = np.random.random(100) 4 [2] 5 L 6 array([0.3516844 , 0.65757053, 0.92102817, 0.08232176, 0.29381881, 7 0.65003014, 0.97971106, 0.5226598 , 0.30206059, 0.08943851, 8 0.18012353, 0.73187738, 0.21161704, 0.99245898, 0.72893738, 9 0.81803006, 0.46057643, 0.78305203, 0.21144964, 0.36252883, 10 0.35272908, 0.25604162, 0.28917731, 0.67740027, 0.90426583, 11 0.84544952, 0.40110828, 0.04247786, 0.01085229, 0.34014734, 12 0.70602233, 0.51424636, 0.39786138, 0.36595197, 0.07393888, 13 0.75163777, 0.45510184, 0.08657692, 0.67168446, 0.88948244, 14 0.2225157 , 0.2878433 , 0.49303834, 0.81036448, 0.01278707, 15 0.10601227, 0.89256786, 0.44614657, 0.43688635, 0.79290611, 16 0.74071001, 0.88761407, 0.28543688, 0.29216674, 0.11519011, 17 0.73617155, 0.3164547 , 0.37765957, 0.02654595, 0.54816316, 18 0.81750059, 0.63684799, 0.27424713, 0.1936451 , 0.06800676, 19 0.8512259 , 0.9296003 , 0.32355494, 0.85864702, 0.59670948, 20 0.34124613, 0.31258873, 0.64220369, 0.14140289, 0.61121421, 21 0.05605693, 0.79119813, 0.27604157, 0.74071167, 0.81422001, 22 0.68846038, 0.56658745, 0.75986556, 0.84020345, 0.93567278, 23 0.20082763, 0.77921029, 0.22234662, 0.63349144, 0.91545638, 24 0.44481323, 0.89917595, 0.37389024, 0.68454439, 0.13811925, 25 0.62240377, 0.96188648, 0.1028726 , 0.11962078, 0.91129131]) 26 [3] 27 sum (L) 28 50.26791878652218 29 [4] 30 np.sum(L) 31 50.267918786522195 32 [5] 33 big_array = np.random.rand(1000000) 34 %timeit sum(big_array) 35 %timeit np.sum(big_array) 36 151 ms ± 2.28 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) 37 1.96 ms ± 239 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each) 38 39 [59] 40 big_array 41 array([0.02665502, 0.79968067, 0.15228547, ..., 0.26373403, 0.7936674 , 42 0.82165722]) 43 [6] 44 np.min(big_array) 45 2.3131487036920362e-07 46 [7] 47 np.max(big_array) 48 0.9999998477089909 49 [8] 50 big_array.min() 51 2.3131487036920362e-07 52 [9] 53 big_array.max() 54 0.9999998477089909 55 [10] 56 big_array.sum() 57 499562.4579554482 58 [11] 59 X = np.arange(16).reshape(4,-1) 60 X 61 array([[ 0, 1, 2, 3], 62 [ 4, 5, 6, 7], 63 [ 8, 9, 10, 11], 64 [12, 13, 14, 15]]) 65 [12] 66 np.sum(X) 67 120 68 [13] 69 np.sum(X,axis = 0) 70 array([24, 28, 32, 36]) 71 [14] 72 np.sum(X,axis = 1) 73 array([ 6, 22, 38, 54]) 74 [15] 75 np.prod(X) # 返回给定维度上各个元素的乘积 76 0 77 [16] 78 np.prod(X + 1) 79 2004189184 80 [17] 81 np.mean(X) 82 7.5 83 [18] 84 np.median(X) 85 7.5 86 [19] 87 v = np.array([1,1,2,2,10]) 88 np.mean(v) 89 3.2 90 [20] 91 np.median(v) 92 2.0 93 [21] 94 np.percentile(big_array,q = 50) # 有50%小于它的数 95 0.49877343789608786 96 [22] 97 np.median(big_array) 98 0.49877343789608786 99 [23] 100 np.percentile(big_array,q = 100) 101 0.9999998477089909 102 [24] 103 np.max(big_array) 104 0.9999998477089909 105 [25] 106 for percent in [0,25,50,75,100]: 107 print(np.percentile(big_array,q = percent)) 108 2.3131487036920362e-07 109 0.24949869046558878 110 0.49877343789608786 111 0.7495843236030313 112 0.9999998477089909 113 114 [26] 115 np.var(big_array)# 方差 116 0.08337201925355782 117 [27] 118 np.std(big_array) 119 0.28874213279941985 120 [28] 121 x = np.random.normal(0,1,size = 1000000) 122 [29] 123 np.mean(x) 124 -0.0014305613307867217 125 [30] 126 np.std(x) 127 1.00021235447686 128 索引 129 [31] 130 np.min(x) 131 -4.691580943262158 132 [32] 133 np.argmin(x) 134 858651 135 [33] 136 x[989892] 137 -0.6379807273320142 138 [34] 139 np.argmax(x) 140 721252 141 [35] 142 x[215038] 143 -2.267033547724769 144 [36] 145 np.max(x) 146 4.710557431454966 147 排序和使用索引 148 [37] 149 x = np.arange(16) 150 x 151 array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) 152 [38] 153 np.random.shuffle(x) 154 x 155 array([ 1, 13, 0, 2, 15, 7, 10, 5, 3, 11, 4, 12, 6, 9, 8, 14]) 156 [39] 157 np.sort(x) 158 array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) 159 [40] 160 x 161 array([ 1, 13, 0, 2, 15, 7, 10, 5, 3, 11, 4, 12, 6, 9, 8, 14]) 162 [41] 163 x.sort() 164 [42] 165 x 166 array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) 167 [43] 168 X = np.random.randint(10,size = (4,4)) 169 X 170 array([[0, 0, 2, 9], 171 [0, 2, 3, 2], 172 [9, 1, 3, 4], 173 [6, 8, 0, 2]]) 174 [44] 175 np.sort(X) 176 array([[0, 0, 2, 9], 177 [0, 2, 2, 3], 178 [1, 3, 4, 9], 179 [0, 2, 6, 8]]) 180 [45] 181 np.sort(X,axis = 1) 182 array([[0, 0, 2, 9], 183 [0, 2, 2, 3], 184 [1, 3, 4, 9], 185 [0, 2, 6, 8]]) 186 [46] 187 np.sort(X,axis = 0 ) 188 array([[0, 0, 0, 2], 189 [0, 1, 2, 2], 190 [6, 2, 3, 4], 191 [9, 8, 3, 9]]) 192 [47] 193 x 194 array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) 195 [48] 196 np.random.shuffle(x) 197 [49] 198 x 199 array([12, 14, 5, 11, 4, 10, 7, 3, 2, 0, 15, 13, 6, 9, 1, 8]) 200 [50] 201 np.argsort(x) 202 array([ 9, 14, 8, 7, 4, 2, 12, 6, 15, 13, 5, 3, 0, 11, 1, 10], 203 dtype=int64) 204 [51] 205 np.partition(x,3) 206 array([ 0, 1, 2, 3, 4, 6, 7, 5, 8, 14, 15, 13, 10, 9, 11, 12]) 207 [52] 208 np.argpartition(x,3) 209 array([ 9, 14, 8, 7, 4, 12, 6, 2, 15, 1, 10, 11, 5, 13, 3, 0], 210 dtype=int64) 211 [53] 212 X 213 array([[0, 0, 2, 9], 214 [0, 2, 3, 2], 215 [9, 1, 3, 4], 216 [6, 8, 0, 2]]) 217 [54] 218 np.argsort(X,axis = 1) 219 array([[0, 1, 2, 3], 220 [0, 1, 3, 2], 221 [1, 2, 3, 0], 222 [2, 3, 0, 1]], dtype=int64) 223 [55] 224 np.argsort(X,axis = 0) 225 array([[0, 0, 3, 1], 226 [1, 2, 0, 3], 227 [3, 1, 1, 2], 228 [2, 3, 2, 0]], dtype=int64) 229 [56] 230 np.argsort(X) 231 array([[0, 1, 2, 3], 232 [0, 1, 3, 2], 233 [1, 2, 3, 0], 234 [2, 3, 0, 1]], dtype=int64) 235 [61] 236 np.partition(X,2) 237 array([[0, 0, 2, 9], 238 [0, 2, 2, 3], 239 [1, 3, 4, 9], 240 [0, 2, 6, 8]]) 241 [57] 242 np.argpartition(X,2,axis = 1) 243 array([[0, 1, 2, 3], 244 [0, 1, 3, 2], 245 [1, 2, 3, 0], 246 [2, 3, 0, 1]], dtype=int64) 247 [58] 248 np.argpartition(X,2,axis = 0) 249 array([[0, 0, 3, 1], 250 [1, 2, 0, 3], 251 [3, 1, 2, 2], 252 [2, 3, 1, 0]], dtype=int64)
3-10 Numpy中的比较和FancyIndexing
Notbook 示例
Notbook 源码
1 fancy indexing 2 [1] 3 import numpy as np 4 5 x = np.arange(16) 6 x 7 array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) 8 [2] 9 x[3] 10 3 11 [3] 12 x[3:9] 13 array([3, 4, 5, 6, 7, 8]) 14 [4] 15 x[3:9:2] 16 array([3, 5, 7]) 17 [5] 18 [x[3],x[5],x[8]] 19 [3, 5, 8] 20 [6] 21 ind = [3,5,8] 22 [7] 23 x[ind] 24 array([3, 5, 8]) 25 用inw也行，只要对上就OK 26 27 [8] 28 ind = np.array([[0,2], 29 [1,3]]) 30 x[ind] 31 array([[0, 2], 32 [1, 3]]) 33 [9] 34 X = x.reshape(4,-1) 35 X 36 array([[ 0, 1, 2, 3], 37 [ 4, 5, 6, 7], 38 [ 8, 9, 10, 11], 39 [12, 13, 14, 15]]) 40 [10] 41 row = np.array([0,1,2]) 42 col = np.array([1,2,3]) 43 X[row,col] 44 array([ 1, 6, 11]) 45 [11] 46 X[0,col] 47 array([1, 2, 3]) 48 [12] 49 X [:2,col] 50 array([[1, 2, 3], 51 [5, 6, 7]]) 52 [13] 53 col = [True ,False,True ,True] 54 [43] 55 X = x.reshape(4,-1) 56 [46] 57 X[1:3,col] 58 array([[ 4, 6, 7], 59 [ 8, 10, 11]]) 60 numpy array的比较 61 [15] 62 x 63 array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) 64 [16] 65 x < 3 66 array([ True, True, True, False, False, False, False, False, False, 67 False, False, False, False, False, False, False]) 68 [17] 69 x > 3 70 array([False, False, False, False, True, True, True, True, True, 71 True, True, True, True, True, True, True]) 72 [18] 73 x <= 3 74 array([ True, True, True, True, False, False, False, False, False, 75 False, False, False, False, False, False, False]) 76 [19] 77 x >= 3 78 array([False, False, False, True, True, True, True, True, True, 79 True, True, True, True, True, True, True]) 80 [20] 81 x == 3 82 array([False, False, False, True, False, False, False, False, False, 83 False, False, False, False, False, False, False]) 84 [21] 85 x != 3 86 array([ True, True, True, False, True, True, True, True, True, 87 True, True, True, True, True, True, True]) 88 [22] 89 2 * x == 24 - 4*x 90 array([False, False, False, False, True, False, False, False, False, 91 False, False, False, False, False, False, False]) 92 [23] 93 X 94 array([[ 0, 1, 2, 3], 95 [ 4, 5, 6, 7], 96 [ 8, 9, 10, 11], 97 [12, 13, 14, 15]]) 98 [24] 99 X < 6 100 array([[ True, True, True, True], 101 [ True, True, False, False], 102 [False, False, False, False], 103 [False, False, False, False]]) 104 [25] 105 x 106 array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) 107 [26] 108 np.sum(x <= 3) 109 4 110 [27] 111 np.count_nonzero(x <= 3) 112 4 113 [28] 114 np.any(x == 0) #只要有一个元素满足条件则True 115 True 116 [29] 117 np.any( x<0 ) 118 False 119 [30] 120 np.all( x >= 0 ) 121 True 122 [31] 123 np.all( x > 0) 124 False 125 [32] 126 X 127 array([[ 0, 1, 2, 3], 128 [ 4, 5, 6, 7], 129 [ 8, 9, 10, 11], 130 [12, 13, 14, 15]]) 131 [33] 132 np.sum( X % 2 == 0) 133 8 134 [34] 135 np.sum( X % 2 == 0 , axis = 1) 136 array([2, 2, 2, 2]) 137 [35] 138 np.sum( X % 2 == 0 , axis = 0) 139 array([4, 0, 4, 0]) 140 [36] 141 np.all(X > 0 ,axis = 1) 142 array([False, True, True, True]) 143 [37] 144 np.sum((x > 3) & (x < 10)) #用&&则报错 145 6 146 [38] 147 np.sum((x%2==0) | (x>10)) 148 11 149 [39] 150 np.sum(~(x==0)) 151 15 152 [40] 153 x[x<5] 154 array([0, 1, 2, 3, 4]) 155 [41] 156 x[x%2==0] 157 array([ 0, 2, 4, 6, 8, 10, 12, 14]) 158 [47] 159 X[ X[:,3] % 3 == 0,:] 160 # 布尔值决定取舍 161 #即，X [[1,0,0,1],:] 162 array([[ 0, 1, 2, 3], 163 [12, 13, 14, 15]])
3-11 Matplotlib数据可视化基础
Notbook 示例
Notbook 源码
1 [1]
2 #import matplotlib as mpl
3 [2]
4 import matplotlib.pyplot as plt
5 [3]
6 import numpy as np
7 [4]
8 x = np.linspace(0,10,100)
9 [5]
10 x
11 array([ 0. , 0.1010101 , 0.2020202 , 0.3030303 , 0.4040404 ,
12 0.50505051, 0.60606061, 0.70707071, 0.80808081, 0.90909091,
13 1.01010101, 1.11111111, 1.21212121, 1.31313131, 1.41414141,
14 1.51515152, 1.61616162, 1.71717172, 1.81818182, 1.91919192,
15 2.02020202, 2.12121212, 2.22222222, 2.32323232, 2.42424242,
16 2.52525253, 2.62626263, 2.72727273, 2.82828283, 2.92929293,
17 3.03030303, 3.13131313, 3.23232323, 3.33333333, 3.43434343,
18 3.53535354, 3.63636364, 3.73737374, 3.83838384, 3.93939394,
19 4.04040404, 4.14141414, 4.24242424, 4.34343434, 4.44444444,
20 4.54545455, 4.64646465, 4.74747475, 4.84848485, 4.94949495,
21 5.05050505, 5.15151515, 5.25252525, 5.35353535, 5.45454545,
22 5.55555556, 5.65656566, 5.75757576, 5.85858586, 5.95959596,
23 6.06060606, 6.16161616, 6.26262626, 6.36363636, 6.46464646,
24 6.56565657, 6.66666667, 6.76767677, 6.86868687, 6.96969697,
25 7.07070707, 7.17171717, 7.27272727, 7.37373737, 7.47474747,
26 7.57575758, 7.67676768, 7.77777778, 7.87878788, 7.97979798,
27 8.08080808, 8.18181818, 8.28282828, 8.38383838, 8.48484848,
28 8.58585859, 8.68686869, 8.78787879, 8.88888889, 8.98989899,
29 9.09090909, 9.19191919, 9.29292929, 9.39393939, 9.49494949,
30 9.5959596 , 9.6969697 , 9.7979798 , 9.8989899 , 10. ])
31 [6]
32 y = np.sin(x)
33 [7]
34 y
35 array([ 0. , 0.10083842, 0.20064886, 0.2984138 , 0.39313661,
36 0.48385164, 0.56963411, 0.64960951, 0.72296256, 0.78894546,
37 0.84688556, 0.8961922 , 0.93636273, 0.96698762, 0.98775469,
38 0.99845223, 0.99897117, 0.98930624, 0.96955595, 0.93992165,
39 0.90070545, 0.85230712, 0.79522006, 0.73002623, 0.65739025,
40 0.57805259, 0.49282204, 0.40256749, 0.30820902, 0.21070855,
41 0.11106004, 0.01027934, -0.09060615, -0.19056796, -0.28858706,
42 -0.38366419, -0.47483011, -0.56115544, -0.64176014, -0.7158225 ,
43 -0.7825875 , -0.84137452, -0.89158426, -0.93270486, -0.96431712,
44 -0.98609877, -0.99782778, -0.99938456, -0.99075324, -0.97202182,
45 -0.94338126, -0.90512352, -0.85763861, -0.80141062, -0.73701276,
46 -0.66510151, -0.58640998, -0.50174037, -0.41195583, -0.31797166,
47 -0.22074597, -0.12126992, -0.0205576 , 0.0803643 , 0.18046693,
48 0.27872982, 0.37415123, 0.46575841, 0.55261747, 0.63384295,
49 0.7086068 , 0.77614685, 0.83577457, 0.8868821 , 0.92894843,
50 0.96154471, 0.98433866, 0.99709789, 0.99969234, 0.99209556,
51 0.97438499, 0.94674118, 0.90944594, 0.86287948, 0.8075165 ,
52 0.74392141, 0.6727425 , 0.59470541, 0.51060568, 0.42130064,
53 0.32770071, 0.23076008, 0.13146699, 0.03083368, -0.07011396,
54 -0.17034683, -0.26884313, -0.36459873, -0.45663749, -0.54402111])
55 [8]
56 plt.plot(x,y)# 不用plt.show()直接出图
57 [<matplotlib.lines.Line2D at 0x1b74c435760>]
58
59 [9]
60 cosy = np.cos(x)
61 [10]
62 cosy.shape
63 (100,)
64 [11]
65 siny = y.copy()
66 [12]
67 plt.plot(x,siny)
68 plt.plot(x,cosy)
69 [<matplotlib.lines.Line2D at 0x1b74c539b80>]
70
71 [13]
72 plt.plot(x,siny)
73 plt.plot(x,cosy,color = 'red')
74 [<matplotlib.lines.Line2D at 0x1b74c5b2700>]
75
76 [14]
77 plt.plot(x,siny)
78 plt.plot(x,cosy,color = 'red',linestyle = '--')
79 [<matplotlib.lines.Line2D at 0x1b74c61adf0>]
80
81 [15]
82 plt.plot(x,siny)
83 plt.plot(x,cosy,color = 'red',linestyle = '--')
84 plt.xlim(-5,15)
85 (-5.0, 15.0)
86
87 [16]
88 plt.plot(x,siny)
89 plt.plot(x,cosy,color = 'red',linestyle = '--')
90 plt.xlim(-5,15)
91 plt.ylim(0,1.5)
92 (0.0, 1.5)
93
94 [17]
95 plt.plot(x,siny)
96 plt.plot(x,cosy,color = 'red',linestyle = '--')
97 plt.axis([-1,11,-2,2])
98 (-1.0, 11.0, -2.0, 2.0)
99
100 [18]
101 plt.plot(x,siny)
102 plt.plot(x,cosy,color = 'red',linestyle = '--')
103 plt.xlabel("x label")
104 plt.ylabel('y value')
105 Text(0, 0.5, 'y value')
106
107 [19]
108 plt.plot(x,siny,label= 'sin(x)')
109 plt.plot(x,cosy,color = 'red',linestyle = '--',label= 'cos(x)')
110 plt.xlabel("x label")
111 plt.ylabel('y value')
112 plt.legend() # 无sin(x)与cos(x)下标,legend(),需要加括号
113 <matplotlib.legend.Legend at 0x1b74d815bb0>
114
115 [20]
116 plt.plot(x,siny,label= 'sin(x)')
117 plt.plot(x,cosy,color = 'red',linestyle = '--',label= 'cos(x)')
118 plt.xlabel("x label")
119 plt.ylabel('y value')
120 plt.legend()
121 plt.title("welcom to ML world")
122 Text(0.5, 1.0, 'welcom to ML world')
123
124 scatter plot
125 [21]
126 plt.scatter(x,siny)
127 <matplotlib.collections.PathCollection at 0x1b74d9064f0>
128
129 [22]
130 plt.scatter(x,siny)
131 plt.scatter(x,cosy,color = "red")
132 <matplotlib.collections.PathCollection at 0x1b74c764a60>
133
134 [23]
135 x = np.random.normal(0,1,10000)
136 y = np.random.normal(0,1,10000)
137 plt.scatter(x,y,alpha = 0.5)
138 <matplotlib.collections.PathCollection at 0x1b74c6eac10>
3-12 数据加载和简单的数据探索
Notbook 示例
Notbook 源码
1 1] 2 import numpy as np 3 [2] 4 # import matplotlib as mpl 5 import matplotlib.pyplot as plt 6 [3] 7 from sklearn import datasets 8 [4] 9 iris = datasets.load_iris() 10 [5] 11 iris.keys() 12 dict_keys(['data', 'target', 'frame', 'target_names', 'DESCR', 'feature_names', 'filename', 'data_module']) 13 [6] 14 print(iris.DESCR) 15 .. _iris_dataset: 16 17 Iris plants dataset 18 -------------------- 19 20 **Data Set Characteristics:** 21 22 :Number of Instances: 150 (50 in each of three classes) 23 :Number of Attributes: 4 numeric, predictive attributes and the class 24 :Attribute Information: 25 - sepal length in cm 26 - sepal width in cm 27 - petal length in cm 28 - petal width in cm 29 - class: 30 - Iris-Setosa 31 - Iris-Versicolour 32 - Iris-Virginica 33 34 :Summary Statistics: 35 36 ============== ==== ==== ======= ===== ==================== 37 Min Max Mean SD Class Correlation 38 ============== ==== ==== ======= ===== ==================== 39 sepal length: 4.3 7.9 5.84 0.83 0.7826 40 sepal width: 2.0 4.4 3.05 0.43 -0.4194 41 petal length: 1.0 6.9 3.76 1.76 0.9490 (high!) 42 petal width: 0.1 2.5 1.20 0.76 0.9565 (high!) 43 ============== ==== ==== ======= ===== ==================== 44 45 :Missing Attribute Values: None 46 :Class Distribution: 33.3% for each of 3 classes. 47 :Creator: R.A. Fisher 48 :Donor: Michael Marshall (MARSHALL%[email protected]) 49 :Date: July, 1988 50 51 The famous Iris database, first used by Sir R.A. Fisher. The dataset is taken 52 from Fisher's paper. Note that it's the same as in R, but not as in the UCI 53 Machine Learning Repository, which has two wrong data points. 54 55 This is perhaps the best known database to be found in the 56 pattern recognition literature. Fisher's paper is a classic in the field and 57 is referenced frequently to this day. (See Duda & Hart, for example.) The 58 data set contains 3 classes of 50 instances each, where each class refers to a 59 type of iris plant. One class is linearly separable from the other 2; the 60 latter are NOT linearly separable from each other. 61 62 .. topic:: References 63 64 - Fisher, R.A. "The use of multiple measurements in taxonomic problems" 65 Annual Eugenics, 7, Part II, 179-188 (1936); also in "Contributions to 66 Mathematical Statistics" (John Wiley, NY, 1950). 67 - Duda, R.O., & Hart, P.E. (1973) Pattern Classification and Scene Analysis. 68 (Q327.D83) John Wiley & Sons. ISBN 0-471-22361-1. See page 218. 69 - Dasarathy, B.V. (1980) "Nosing Around the Neighborhood: A New System 70 Structure and Classification Rule for Recognition in Partially Exposed 71 Environments". IEEE Transactions on Pattern Analysis and Machine 72 Intelligence, Vol. PAMI-2, No. 1, 67-71. 73 - Gates, G.W. (1972) "The Reduced Nearest Neighbor Rule". IEEE Transactions 74 on Information Theory, May 1972, 431-433. 75 - See also: 1988 MLC Proceedings, 54-64. Cheeseman et al"s AUTOCLASS II 76 conceptual clustering system finds 3 classes in the data. 77 - Many, many more ... 78 79 [7] 80 iris.data 81 array([[5.1, 3.5, 1.4, 0.2], 82 [4.9, 3. , 1.4, 0.2], 83 [4.7, 3.2, 1.3, 0.2], 84 [4.6, 3.1, 1.5, 0.2], 85 [5. , 3.6, 1.4, 0.2], 86 [5.4, 3.9, 1.7, 0.4], 87 [4.6, 3.4, 1.4, 0.3], 88 [5. , 3.4, 1.5, 0.2], 89 [4.4, 2.9, 1.4, 0.2], 90 [4.9, 3.1, 1.5, 0.1], 91 [5.4, 3.7, 1.5, 0.2], 92 [4.8, 3.4, 1.6, 0.2], 93 [4.8, 3. , 1.4, 0.1], 94 [4.3, 3. , 1.1, 0.1], 95 [5.8, 4. , 1.2, 0.2], 96 [5.7, 4.4, 1.5, 0.4], 97 [5.4, 3.9, 1.3, 0.4], 98 [5.1, 3.5, 1.4, 0.3], 99 [5.7, 3.8, 1.7, 0.3], 100 [5.1, 3.8, 1.5, 0.3], 101 [5.4, 3.4, 1.7, 0.2], 102 [5.1, 3.7, 1.5, 0.4], 103 [4.6, 3.6, 1. , 0.2], 104 [5.1, 3.3, 1.7, 0.5], 105 [4.8, 3.4, 1.9, 0.2], 106 [5. , 3. , 1.6, 0.2], 107 [5. , 3.4, 1.6, 0.4], 108 [5.2, 3.5, 1.5, 0.2], 109 [5.2, 3.4, 1.4, 0.2], 110 [4.7, 3.2, 1.6, 0.2], 111 [4.8, 3.1, 1.6, 0.2], 112 [5.4, 3.4, 1.5, 0.4], 113 [5.2, 4.1, 1.5, 0.1], 114 [5.5, 4.2, 1.4, 0.2], 115 [4.9, 3.1, 1.5, 0.2], 116 [5. , 3.2, 1.2, 0.2], 117 [5.5, 3.5, 1.3, 0.2], 118 [4.9, 3.6, 1.4, 0.1], 119 [4.4, 3. , 1.3, 0.2], 120 [5.1, 3.4, 1.5, 0.2], 121 [5. , 3.5, 1.3, 0.3], 122 [4.5, 2.3, 1.3, 0.3], 123 [4.4, 3.2, 1.3, 0.2], 124 [5. , 3.5, 1.6, 0.6], 125 [5.1, 3.8, 1.9, 0.4], 126 [4.8, 3. , 1.4, 0.3], 127 [5.1, 3.8, 1.6, 0.2], 128 [4.6, 3.2, 1.4, 0.2], 129 [5.3, 3.7, 1.5, 0.2], 130 [5. , 3.3, 1.4, 0.2], 131 [7. , 3.2, 4.7, 1.4], 132 [6.4, 3.2, 4.5, 1.5], 133 [6.9, 3.1, 4.9, 1.5], 134 [5.5, 2.3, 4. , 1.3], 135 [6.5, 2.8, 4.6, 1.5], 136 [5.7, 2.8, 4.5, 1.3], 137 [6.3, 3.3, 4.7, 1.6], 138 [4.9, 2.4, 3.3, 1. ], 139 [6.6, 2.9, 4.6, 1.3], 140 [5.2, 2.7, 3.9, 1.4], 141 [5. , 2. , 3.5, 1. ], 142 [5.9, 3. , 4.2, 1.5], 143 [6. , 2.2, 4. , 1. ], 144 [6.1, 2.9, 4.7, 1.4], 145 [5.6, 2.9, 3.6, 1.3], 146 [6.7, 3.1, 4.4, 1.4], 147 [5.6, 3. , 4.5, 1.5], 148 [5.8, 2.7, 4.1, 1. ], 149 [6.2, 2.2, 4.5, 1.5], 150 [5.6, 2.5, 3.9, 1.1], 151 [5.9, 3.2, 4.8, 1.8], 152 [6.1, 2.8, 4. , 1.3], 153 [6.3, 2.5, 4.9, 1.5], 154 [6.1, 2.8, 4.7, 1.2], 155 [6.4, 2.9, 4.3, 1.3], 156 [6.6, 3. , 4.4, 1.4], 157 [6.8, 2.8, 4.8, 1.4], 158 [6.7, 3. , 5. , 1.7], 159 [6. , 2.9, 4.5, 1.5], 160 [5.7, 2.6, 3.5, 1. ], 161 [5.5, 2.4, 3.8, 1.1], 162 [5.5, 2.4, 3.7, 1. ], 163 [5.8, 2.7, 3.9, 1.2], 164 [6. , 2.7, 5.1, 1.6], 165 [5.4, 3. , 4.5, 1.5], 166 [6. , 3.4, 4.5, 1.6], 167 [6.7, 3.1, 4.7, 1.5], 168 [6.3, 2.3, 4.4, 1.3], 169 [5.6, 3. , 4.1, 1.3], 170 [5.5, 2.5, 4. , 1.3], 171 [5.5, 2.6, 4.4, 1.2], 172 [6.1, 3. , 4.6, 1.4], 173 [5.8, 2.6, 4. , 1.2], 174 [5. , 2.3, 3.3, 1. ], 175 [5.6, 2.7, 4.2, 1.3], 176 [5.7, 3. , 4.2, 1.2], 177 [5.7, 2.9, 4.2, 1.3], 178 [6.2, 2.9, 4.3, 1.3], 179 [5.1, 2.5, 3. , 1.1], 180 [5.7, 2.8, 4.1, 1.3], 181 [6.3, 3.3, 6. , 2.5], 182 [5.8, 2.7, 5.1, 1.9], 183 [7.1, 3. , 5.9, 2.1], 184 [6.3, 2.9, 5.6, 1.8], 185 [6.5, 3. , 5.8, 2.2], 186 [7.6, 3. , 6.6, 2.1], 187 [4.9, 2.5, 4.5, 1.7], 188 [7.3, 2.9, 6.3, 1.8], 189 [6.7, 2.5, 5.8, 1.8], 190 [7.2, 3.6, 6.1, 2.5], 191 [6.5, 3.2, 5.1, 2. ], 192 [6.4, 2.7, 5.3, 1.9], 193 [6.8, 3. , 5.5, 2.1], 194 [5.7, 2.5, 5. , 2. ], 195 [5.8, 2.8, 5.1, 2.4], 196 [6.4, 3.2, 5.3, 2.3], 197 [6.5, 3. , 5.5, 1.8], 198 [7.7, 3.8, 6.7, 2.2], 199 [7.7, 2.6, 6.9, 2.3], 200 [6. , 2.2, 5. , 1.5], 201 [6.9, 3.2, 5.7, 2.3], 202 [5.6, 2.8, 4.9, 2. ], 203 [7.7, 2.8, 6.7, 2. ], 204 [6.3, 2.7, 4.9, 1.8], 205 [6.7, 3.3, 5.7, 2.1], 206 [7.2, 3.2, 6. , 1.8], 207 [6.2, 2.8, 4.8, 1.8], 208 [6.1, 3. , 4.9, 1.8], 209 [6.4, 2.8, 5.6, 2.1], 210 [7.2, 3. , 5.8, 1.6], 211 [7.4, 2.8, 6.1, 1.9], 212 [7.9, 3.8, 6.4, 2. ], 213 [6.4, 2.8, 5.6, 2.2], 214 [6.3, 2.8, 5.1, 1.5], 215 [6.1, 2.6, 5.6, 1.4], 216 [7.7, 3. , 6.1, 2.3], 217 [6.3, 3.4, 5.6, 2.4], 218 [6.4, 3.1, 5.5, 1.8], 219 [6. , 3. , 4.8, 1.8], 220 [6.9, 3.1, 5.4, 2.1], 221 [6.7, 3.1, 5.6, 2.4], 222 [6.9, 3.1, 5.1, 2.3], 223 [5.8, 2.7, 5.1, 1.9], 224 [6.8, 3.2, 5.9, 2.3], 225 [6.7, 3.3, 5.7, 2.5], 226 [6.7, 3. , 5.2, 2.3], 227 [6.3, 2.5, 5. , 1.9], 228 [6.5, 3. , 5.2, 2. ], 229 [6.2, 3.4, 5.4, 2.3], 230 [5.9, 3. , 5.1, 1.8]]) 231 [8] 232 iris.data.shape 233 (150, 4) 234 [9] 235 iris.feature_names 236 ['sepal length (cm)', 237 'sepal width (cm)', 238 'petal length (cm)', 239 'petal width (cm)'] 240 [10] 241 iris.target 242 array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 243 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 244 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 245 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 246 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 247 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 248 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]) 249 [11] 250 iris.target.shape 251 (150,) 252 [12] 253 iris.target_names 254 array(['setosa', 'versicolor', 'virginica'], dtype='<U10') 255 [13] 256 X = iris.data[:,:2] 257 [14] 258 X.shape 259 (150, 2) 260 [15] 261 plt.scatter(X[:,0],X[:,1]) 262 <matplotlib.collections.PathCollection at 0x1aa5b280280> 263 264 [16] 265 y = iris.target 266 [17] 267 plt.scatter(X[y == 0,0],X[y == 0,1],color = 'red') 268 plt.scatter(X[y == 1,0],X[y == 1,1],color = 'blue') 269 plt.scatter(X[y == 2,0],X[y == 2,1],color = 'green') 270 <matplotlib.collections.PathCollection at 0x1aa5b386d30> 271 272 [18] 273 plt.scatter(X[y == 0,0],X[y == 0,1],color = 'red',marker='o') 274 plt.scatter(X[y == 1,0],X[y == 1,1],color = 'blue',marker='+') 275 plt.scatter(X[y == 2,0],X[y == 2,1],color = 'green',marker='x') 276 <matplotlib.collections.PathCollection at 0x1aa5b4089d0> 277 278 [19] 279 X = iris.data[:,2:] 280 [20] 281 plt.scatter(X[y == 0,0],X[y == 0,1],color = 'red',marker='o') 282 plt.scatter(X[y == 1,0],X[y == 1,1],color = 'blue',marker='+') 283 plt.scatter(X[y == 2,0],X[y == 2,1],color = 'green',marker='x') 284 <matplotlib.collections.PathCollection at 0x1aa5b46f640>
标签：11,10,12,Jupyter,666,Notebook,np,array,numpy From： https://www.cnblogs.com/Cai-Gbro/p/16839828.html