資源簡介
Discuss the influence of the following factors (plot the
results and quantify the test error and the training error):
1. # of training samples
代碼片段和文件信息
clc;clear;clf;
load?data_D2_C2.mat
%%
%?[DNp]=size(p.value);
%?for?i=1:Np
%?????if?p.class(i)==1
%?????????plot(p.value(1i)p.value(2i)‘r.‘);hold?on;
%?????else
%?????????plot(p.value(1i)p.value(2i)‘b.‘);hold?on;
%?????end??
%?end
[DNt]=size(t.value);
for?i=1:Nt
????if?t.class(i)==1
????????plot(t.value(1i)t.value(2i)‘r.‘);hold?on;
????else
????????plot(t.value(1i)t.value(2i)‘b.‘);hold?on;
????end??
end
no_neurons1=8;
no_neurons2=8;
total_neurons=no_neurons1*no_neurons2;
samples=10:10:300;
for?n=1:30
????????ind_rand=randperm?(300);??ind?=ind_rand(1:samples(n));
????????train_data=zeros(2samples(n));
????????train_class=zeros(1samples(n));
????????for?a=1:samples(n)
????????????train_data(:a)=p.value(:ind(a));
????????????train_class(a)=p.class(:ind(a));
????????end
????????
????????som1=selforgmap([no_neurons1?no_neurons2]);
????????som1=train(som1train_data);
????????plotsom(som1.iw{11}som1.layers{1}.distances);
????????ynp=sim(som1train_data);
????????ynpind=vec2ind(ynp);
n1=zeros(1total_neurons);
n2=zeros(1total_neurons);
for?i=1:samples(n)
????if?train_class(i)==1
????????for?j=1:total_neurons
????????????if?ynpind(i)==j
????????????????n1(1j)=n1(1j)+1;
????????????end???
????????end
????else
????????for?j=1:total_neurons
????????????if?ynpind(i)==j
????????????????n2(1j)=n2(1j)+1;
????????????end
????????end
????end
end?
%observe?the?va
- 上一篇:小型火箭Matlab求解
- 下一篇:QPSK調制解調的MATLAB程序仿真
評論
共有 條評論