1 %% Perceptron Regression
2 close all
3 clear
4
5 %%load data
6 x = load('ex4x.dat');
7 y = load('ex4y.dat');
8
9 x=ones(80,2);
10 for i=1:80
11 x(i,1)=mod(i,10);
12 x(i,2)=floor(i/10);
13 end
14
15 for i=1:80
16 if (x(i,1)+x(i,2))<10
17 y(i)=0;
18 else
19 y(i)=1;
20 end
21 end
22
23 [m, n] = size(x);
24
25 % Add intercept term to x
26 x = [ones(m, 1), x];
27
28 %%draw picture
29 % find returns the indices of the
30 % rows meeting the specified condition
31 pos = find(y == 1);
32 neg = find(y == 0);
33 % Assume the features are in the 2nd and 3rd
34 % columns of x
35 figure('NumberTitle', 'off', 'Name', '感知机');
36 plot(x(pos, 2), x(pos,3), '+');
37 hold on;
38 plot(x(neg, 2), x(neg, 3), 'o');
39
40 %进行初始化
41 s = 1; % 标识符，当s=0时，表示迭代终止
42 n = 0; % 表示迭代的次数
43 N = 10000; %定义N为最大分类判别次数，判别次数超过此值则认定样本无法分类。
44 w= [0,0,0]'; % 取权向量初始值
45
46 % 开始迭代
47 while s
48 J = 0; % 假设初始的分类全部正确
49 for i = 1:size(pos)
50 if (x(pos(i),:)*w)<=0 % 查看x1分类是否错误，在x属于w1却被错误分类的情况下，w'x<0
51 w = w+x(pos(i),:)';% 分类错误，对权向量进行修正
52 J = 1; % 置错误标志位为1
53 end
54 end
55 for i = 1:size(neg)
56 if (x(neg(i),:)*w)>=0 % 查看x2分类是否错误，在x属于w2却被错误分类的情况下，w'x>0
57 w = w-x(neg(i),:)';% 分类错误，对权向量进行修正
58 J = 1; % 置错误标志位为1
59 end
60 end
61 if J==0 % 代价为0，即分类均正确
62 s = 0; % 终止迭代
63 end
64 n = n+1; % 迭代次数加1
65 if n == N
66 s=0;
67 end
68 end
69
70 w=[1;w(2)/w(1);w(3)/w(1)];
71
72 %%Calculate the decision boundary line
73 plot_x = [min(x(:,2)), max(x(:,2))];
74 plot_y = (-1./w(3)).*(w(2).*plot_x +w(1));
75 plot(plot_x, plot_y)
76 legend('Admitted', 'Not admitted', 'Decision Boundary')
77 hold off
