LVQ神经网络的分类

<span style="font-size:18px;">%% 清空环境变量
clear all
clc
warning off
%% 导入数据
load data.mat
a=randperm(569);
Train=data(a(1:500),:);
Test=data(a(501:end),:);
% 训练数据
P_train=Train(:,3:end)';
Tc_train=Train(:,2)';
T_train=ind2vec(Tc_train);
% 测试数据
P_test=Test(:,3:end)';
Tc_test=Test(:,2)';
%% 创建网络
count_B=length(find(Tc_train==1));
count_M=length(find(Tc_train==2));
rate_B=count_B/500;
rate_M=count_M/500;
net=newlvq(minmax(P_train),20,[rate_B rate_M],0.01,'learnlv1');
% 设置网络参数
net.trainParam.epochs=1000;
net.trainParam.show=10;
net.trainParam.lr=0.1;
net.trainParam.goal=0.1;
%% 训练网络
net=train(net,P_train,T_train);
%% 仿真测试
T_sim=sim(net,P_test);
Tc_sim=vec2ind(T_sim);
result=[Tc_sim;Tc_test]
%% 结果显示
total_B=length(find(data(:,2)==1));
total_M=length(find(data(:,2)==2));
number_B=length(find(Tc_test==1));
number_M=length(find(Tc_test==2));
number_B_sim=length(find(Tc_sim==1 & Tc_test==1));
number_M_sim=length(find(Tc_sim==2 &Tc_test==2));
disp(['病例总数：' num2str(569)...
      '  良性：' num2str(total_B)...
      '  恶性：' num2str(total_M)]);
disp(['训练集病例总数：' num2str(500)...
      '  良性：' num2str(count_B)...
      '  恶性：' num2str(count_M)]);
disp(['测试集病例总数：' num2str(69)...
      '  良性：' num2str(number_B)...
      '  恶性：' num2str(number_M)]);
disp(['良性乳腺肿瘤确诊：' num2str(number_B_sim)...
      '  误诊：' num2str(number_B-number_B_sim)...
      '  确诊率p1=' num2str(number_B_sim/number_B*100) '%']);
disp(['恶性乳腺肿瘤确诊：' num2str(number_M_sim)...
      '  误诊：' num2str(number_M-number_M_sim)...
      '  确诊率p2=' num2str(number_M_sim/number_M*100) '%']);
</span>