Scribd
Upload
62 views6 pages

Neural 123

The document discusses using the perceptron model to create a decision boundary to classify points. It then discusses the Adaline model and backpropagation network. It shows code examples for training these models and evaluating their performance.

Uploaded by

api-3764708
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
62 views6 pages

Neural 123

The document discusses using the perceptron model to create a decision boundary to classify points. It then discusses the Adaline model and backpropagation network. It shows code examples for training these models and evaluating their performance.

Uploaded by

api-3764708
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Nguyễn Thành Chung ðKTð1_K48 Nhóm 1

Contents

• 1.Dung mo hinh mang Perceptron tao ra duong ranh gioi de phan loai cac diem
sau
• 2.Mang Adline (Bo loc tien doan)
• 3.Mang Backpropagation
• In ra cac gia tri W ,b

1.1.Dung mo hinh mang Perceptron tao ra duong ranh gioi de phan loai
cac diem sau

Cach 1 : Huan luyen theo kieu Train

clear;clc;
p = [[-5;5] [-3;4] [1;6] [3;8] [0;3] [-5;-3] [-2;-1] [1;-7] [2;1]
[6;2]];%input
t = [1 1 1 1 1 0 0 0 0 0];%%Target
PR = newp([-5 3;3 8],1);% tao mang 2 dau vao tam [-5 3][3 8] va 1
noron
PR.trainParam.epochs = 5;PR = train(PR,p,t);%Huan luyen mang
a = sim(PR,p);e=t-a;
plotpv(p,t);hold on;plotpc(PR.IW{1,1},PR.b{1});
% In ra cac gia tri W va b
w=PR.IW{1,1}%Trong so
b=PR.b{1}%tri so phan cuc
TRAINC, Epoch 0/5
TRAINC, Epoch 5/5
TRAINC, Maximum epoch reached.

w =

-2 8

b =

-7

Thí nghiệm HTTM [email protected] page 1 of 6


Nguyễn Thành Chung ðKTð1_K48 Nhóm 1

Cach 2 : Huan luyen theo kieu Adaptic

clear;clc;
p = [[-5;5] [-3;4] [1;6] [3;8] [0;3] [-5;-3] [-2;-1] [1;-7] [2;1]
[6;2]];%input
t = [1 1 1 1 1 0 0 0 0 0];%%Target
PR = newp([-5 3;3 8],1);% tao mang 2 dau vao tam [-5 3][3 8] va 1
noron
plotpv(p,t);
e=1;
linehandle = plotpc(PR.iw{1},PR.b{1});
while(sse(e)) [PR,y,e]=adapt(PR,p,t);
linehandle = plotpc(PR.iw{1},PR.b{1},linehandle)
end;
% In ra cac gia tri W va b
w=PR.IW{1,1}%Trong so
b=PR.b{1}%tri so phan cuc

linehandle =

304.0038

linehandle =

Thí nghiệm HTTM [email protected] page 2 of 6


Nguyễn Thành Chung ðKTð1_K48 Nhóm 1

304.0039

w =

-2 8

b =

-5

2.Mang Adline (Bo loc tien doan)


clear;clc;
t=0.1:0.1:10;
T = exp(-0.3*t).*cos(4*pi*t);%Target
P = exp(-0.3*(t-0.1)).*cos(4*pi*(t-0.1));%input
plot(t,P,'r:',t,T,'b-');legend('P','T');grid on;%Kiem tra

Thí nghiệm HTTM [email protected] page 3 of 6


Nguyễn Thành Chung ðKTð1_K48 Nhóm 1

Ada = newlin([-1 1],1);%tao mang theo yeu cau


Ada.trainParam.goal = 0.07;% Sai so cho phep
Ada.trainParam.show = 10;
Ada.trainParam.epochs = 20;%So buoc huan luyen
Ada = train(Ada,P,T);%huan luyen mang
a = sim(Ada,P);e=T-a;%kiem tra dat yeu cau
TRAINB, Epoch 0/20, MSE 0.0782468/0.07.
TRAINB, Epoch 10/20, MSE 0.0719035/0.07.
TRAINB, Epoch 20/20, MSE 0.0709057/0.07.
TRAINB, Maximum epoch reached.

Thí nghiệm HTTM [email protected] page 4 of 6


Nguyễn Thành Chung ðKTð1_K48 Nhóm 1

3.Mang Backpropagation
clear;clc;
p = [0 1 2 3 4 5 6 7 8 9];%input
t = [0.84 0.91 0.14 -0.77 -0.96 -0.28 0.66 0.99 1 1];%target
ff = newff(minmax(p),[3 1],{'tansig','purelin'},'traingd');%tao
mang theo yeu cau
ff.trainParam.show = 200;ff.trainParam.lr =
0.05;ff.trainParam.epochs = 1000;
ff.trainParam.goal = 0.01;%sai so cho phep
ff = train(ff,p,t);%huan luyen mang
a = sim(ff,p);e=t-a;%kiem tra dat yeu cau
TRAINGD, Epoch 0/1000, MSE 0.502565/0.01, Gradient 1.34781/1e-010
TRAINGD, Epoch 200/1000, MSE 0.0294352/0.01, Gradient 0.0599319/1e-
010
TRAINGD, Epoch 400/1000, MSE 0.0131764/0.01, Gradient 0.0265628/1e-
010
TRAINGD, Epoch 526/1000, MSE 0.00999319/0.01, Gradient 0.0188419/1e-
010
TRAINGD, Performance goal met.

Thí nghiệm HTTM [email protected] page 5 of 6


Nguyễn Thành Chung ðKTð1_K48 Nhóm 1

In ra cac gia tri W ,b


w=ff.IW{1,1}%Trong so
b=ff.b{1}%Do phan cuc

w =

-0.7786
0.8727
-0.5578

b =

8.4179
-4.4347
1.3052

Published with MATLAB® 7.0.4

Thí nghiệm HTTM [email protected] page 6 of 6

You might also like