【独家发布】Machine Learning: An Algorithmic Perspective(using Java)

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

http://seat.massey.ac.nz/personal/s.r.marsland/MLBook.html

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏2 回帖

关键词：perspective Algorithmic Perspectiv Algorithm Learning Java

1. # Code from Chapter 2 of Machine Learning: An Algorithmic Perspective (2nd Edition)
   
2. # by Stephen Marsland (http://stephenmonika.net)
   
3. 
   
4. # You are free to use, change, or redistribute the code in any way you wish for
   
5. # non-commercial purposes, but please maintain the name of the original author.
   
6. # This code comes with no warranty of any kind.
   
7. 
   
8. # Stephen Marsland, 2008, 2014
   
9. 
   
10. # Plots a 1D Gaussian function
    
11. import pylab as pl
    
12. import numpy as np
    
13. 
    
14. gaussian = lambda x: 1/(np.sqrt(2*np.pi)*1.5)*np.exp(-(x-0)**2/(2*(1.5**2)))
    
15. x = np.arange(-5,5,0.01)
    
16. y = gaussian(x)
    
17. pl.ion()
    
18. pl.plot(x,y,'k',linewidth=3)
    
19. pl.xlabel('x')
    
20. pl.ylabel('y(x)')
    
21. pl.axis([-5,5,0,0.3])
    
22. pl.title('Gaussian Function (mean 0, standard deviation 1.5)')
    
23. pl.show()

复制代码

1. # Code from Chapter 3 of Machine Learning: An Algorithmic Perspective (2nd Edition)
   
2. # by Stephen Marsland (http://stephenmonika.net)
   
3. 
   
4. # You are free to use, change, or redistribute the code in any way you wish for
   
5. # non-commercial purposes, but please maintain the name of the original author.
   
6. # This code comes with no warranty of any kind.
   
7. 
   
8. # Stephen Marsland, 2008, 2014
   
9. 
   
10. import numpy as np
    
11. 
    
12. class pcn:
    
13.         """ A basic Perceptron"""
    
14.        
    
15.         def __init__(self,inputs,targets):
    
16.                 """ Constructor """
    
17.                 # Set up network size
    
18.                 if np.ndim(inputs)>1:
    
19.                         self.nIn = np.shape(inputs)[1]
    
20.                 else:
    
21.                         self.nIn = 1
    
22.        
    
23.                 if np.ndim(targets)>1:
    
24.                         self.nOut = np.shape(targets)[1]
    
25.                 else:
    
26.                         self.nOut = 1
    
27. 
    
28.                 self.nData = np.shape(inputs)[0]
    
29.        
    
30.                 # Initialise network
    
31.                 self.weights = np.random.rand(self.nIn+1,self.nOut)*0.1-0.05
    
32. 
    
33.         def pcntrain(self,inputs,targets,eta,nIterations):
    
34.                 """ Train the thing """       
    
35.                 # Add the inputs that match the bias node
    
36.                 inputs = np.concatenate((inputs,-np.ones((self.nData,1))),axis=1)
    
37.                 # Training
    
38.                 change = range(self.nData)
    
39. 
    
40.                 for n in range(nIterations):
    
41.                        
    
42.                         self.activations = self.pcnfwd(inputs);
    
43.                         self.weights -= eta*np.dot(np.transpose(inputs),self.activations-targets)
    
44.                
    
45.                         # Randomise order of inputs
    
46.                         #np.random.shuffle(change)
    
47.                         #inputs = inputs[change,:]
    
48.                         #targets = targets[change,:]
    
49.                        
    
50.                 #return self.weights
    
51. 
    
52.         def pcnfwd(self,inputs):
    
53.                 """ Run the network forward """
    
54.                 # Compute activations
    
55.                 activations =  np.dot(inputs,self.weights)
    
56. 
    
57.                 # Threshold the activations
    
58.                 return np.where(activations>0,1,0)
    
59. 
    
60. 
    
61.         def confmat(self,inputs,targets):
    
62.                 """Confusion matrix"""
    
63. 
    
64.                 # Add the inputs that match the bias node
    
65.                 inputs = np.concatenate((inputs,-np.ones((self.nData,1))),axis=1)
    
66.                
    
67.                 outputs = np.dot(inputs,self.weights)
    
68.        
    
69.                 nClasses = np.shape(targets)[1]
    
70. 
    
71.                 if nClasses==1:
    
72.                         nClasses = 2
    
73.                         outputs = np.where(outputs>0,1,0)
    
74.                 else:
    
75.                         # 1-of-N encoding
    
76.                         outputs = np.argmax(outputs,1)
    
77.                         targets = np.argmax(targets,1)
    
78. 
    
79.                 cm = np.zeros((nClasses,nClasses))
    
80.                 for i in range(nClasses):
    
81.                         for j in range(nClasses):
    
82.                                 cm[i,j] = np.sum(np.where(outputs==i,1,0)*np.where(targets==j,1,0))
    
83. 
    
84.                 print cm
    
85.                 print np.trace(cm)/np.sum(cm)
    
86.                
    
87. def logic():
    
88.         import pcn
    
89.         """ Run AND and XOR logic functions"""
    
90. 
    
91.         a = np.array([[0,0,0],[0,1,0],[1,0,0],[1,1,1]])
    
92.         b = np.array([[0,0,0],[0,1,1],[1,0,1],[1,1,0]])
    
93. 
    
94.         p = pcn.pcn(a[:,0:2],a[:,2:])
    
95.         p.pcntrain(a[:,0:2],a[:,2:],0.25,10)
    
96.         p.confmat(a[:,0:2],a[:,2:])
    
97. 
    
98.         q = pcn.pcn(b[:,0:2],b[:,2:])
    
99.         q.pcntrain(b[:,0:2],b[:,2:],0.25,10)
    
100.         q.confmat(b[:,0:2],b[:,2:])

复制代码