遗传算法优化计算——建模自变量降维Matlab

最新推荐文章于 2023-11-21 13:14:20 发布

Orlando Allen

最新推荐文章于 2023-11-21 13:14:20 发布

阅读量1k

点赞数 3

分类专栏： matlab常用代码文章标签： matlab 开发语言人工智能学习算法

本文链接：https://blog.csdn.net/Allen1862105/article/details/128084373

版权

遗传算法 BP神经网络特征选择优化乳腺肿瘤诊断

关键词由CSDN通过智能技术生成

matlab常用代码专栏收录该内容

18 篇文章 11 订阅

订阅专栏

什么是遗传算法？

遗传算法（Genetic Algorithm，简称GA）起源于对生物系统所进行的计算机模拟研究，是一种随机全局搜索优化方法，它模拟了自然选择和遗传中发生的复制、交叉(crossover)和变异(mutation)等现象，从任一初始种群（Population）出发，通过随机选择、交叉和变异操作，产生一群更适合环境的个体，使群体进化到搜索空间中越来越好的区域，这样一代一代不断繁衍进化，最后收敛到一群最适应环境的个体（Individual），从而求得问题的优质解。

代码详情

1. main.m为主程序，将该文件夹设置为MATLAB当前工作路径，运行即可。

2. fitness.m为输入自变量优化适应度子函数。

3. de_code.m为输入自变量优化编解码子函数。

4. gabpEval.m为BP网络权值和阈值优化适应度子函数。

5. gadecod.m为BP网络权值和阈值优化编解码子函数。

6. data.mat为数据文件。

7. gaot文件夹为GAOT遗传算法工具箱。
安装步骤：File→Set Path→Add with Subfolders→Save

8. 该程序在MATLAB2009a版本下测试通过，个别函数在低版本中不存在或者调用格式有所不同，参照对应版本中的帮助文档修改即可。

main

%% 遗传算法的优化计算——输入自变量降维

%% 清空环境变量
clear all
clc
warning off
%% 声明全局变量
global P_train T_train P_test T_test  mint maxt S s1
S = 30;
s1 = 50;
%% 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)';
T_train = Train(:,2)';
% 测试数据
P_test = Test(:,3:end)';
T_test = Test(:,2)';
% 显示实验条件
total_B = length(find(data(:,2) == 1));
total_M = length(find(data(:,2) == 2));
count_B = length(find(T_train == 1));
count_M = length(find(T_train == 2));
number_B = length(find(T_test == 1));
number_M = length(find(T_test == 2));
disp('实验条件为：');
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)]);
%% 数据归一化
[P_train,minp,maxp,T_train,mint,maxt] = premnmx(P_train,T_train);
P_test = tramnmx(P_test,minp,maxp);
%% 创建单BP网络
t = cputime;
net_bp = newff(minmax(P_train),[s1,1],{'tansig','purelin'},'trainlm');
% 设置训练参数
net_bp.trainParam.epochs = 1000;
net_bp.trainParam.show = 10;
net_bp.trainParam.goal = 0.1;
net_bp.trainParam.lr = 0.1;
net_bp.trainParam.showwindow = 0;
%% 训练单BP网络
net_bp = train(net_bp,P_train,T_train);
%% 仿真测试单BP网络
tn_bp_sim = sim(net_bp,P_test);
% 反归一化
T_bp_sim = postmnmx(tn_bp_sim,mint,maxt);
e = cputime - t;
T_bp_sim(T_bp_sim > 1.5) = 2;
T_bp_sim(T_bp_sim < 1.5) = 1;
result_bp = [T_bp_sim' T_test'];
%% 结果显示（单BP网络）
number_B_sim = length(find(T_bp_sim == 1 & T_test == 1));
number_M_sim = length(find(T_bp_sim == 2 &T_test == 2));
disp('(1)BP网络的测试结果为：');
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) '%']);
disp(['建模时间为：' num2str(e) 's'] );
%% 遗传算法优化
popu = 20;  
bounds = ones(S,1)*[0,1];
% 产生初始种群
initPop = randint(popu,S,[0 1]);
% 计算初始种群适应度
initFit = zeros(popu,1);
for i = 1:size(initPop,1)
    initFit(i) = de_code(initPop(i,:));
end
initPop = [initPop initFit];
gen = 100; 
% 优化计算
[X,EndPop,BPop,Trace] = ga(bounds,'fitness',[],initPop,[1e-6 1 0],'maxGenTerm',...
    gen,'normGeomSelect',0.09,'simpleXover',2,'boundaryMutation',[2 gen 3]);
[m,n] = find(X == 1);
disp(['优化筛选后的输入自变量编号为:' num2str(n)]);
% 绘制适应度函数进化曲线
figure
plot(Trace(:,1),Trace(:,3),'r:')
hold on
plot(Trace(:,1),Trace(:,2),'b')
xlabel('进化代数')
ylabel('适应度函数')
title('适应度函数进化曲线')
legend('平均适应度函数','最佳适应度函数')
xlim([1 gen])
%% 新训练集/测试集数据提取
p_train = zeros(size(n,2),size(T_train,2));
p_test = zeros(size(n,2),size(T_test,2));
for i = 1:length(n)
    p_train(i,:) = P_train(n(i),:);
    p_test(i,:) = P_test(n(i),:);
end
t_train = T_train;
%% 创建优化BP网络
t = cputime;
net_ga = newff(minmax(p_train),[s1,1],{'tansig','purelin'},'trainlm');
% 训练参数设置
net_ga.trainParam.epochs = 1000;
net_ga.trainParam.show = 10;
net_ga.trainParam.goal = 0.1;
net_ga.trainParam.lr = 0.1;
net_ga.trainParam.showwindow = 0;
%% 训练优化BP网络
net_ga = train(net_ga,p_train,t_train);
%% 仿真测试优化BP网络
tn_ga_sim = sim(net_ga,p_test);
% 反归一化
T_ga_sim = postmnmx(tn_ga_sim,mint,maxt);
e = cputime - t;
T_ga_sim(T_ga_sim > 1.5) = 2;
T_ga_sim(T_ga_sim < 1.5) = 1;
result_ga = [T_ga_sim' T_test'];
%% 结果显示（优化BP网络）
number_b_sim = length(find(T_ga_sim == 1 & T_test == 1));
number_m_sim = length(find(T_ga_sim == 2 &T_test == 2));
disp('(2)优化BP网络的测试结果为：');
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) '%']);
disp(['建模时间为：' num2str(e) 's'] );
%% 遗传算法的优化计算——输入自变量降维

%% 清空环境变量
clear all
clc
warning off
%% 声明全局变量
global P_train T_train P_test T_test  mint maxt S s1
S = 30;
s1 = 50;
%% 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)';
T_train = Train(:,2)';
% 测试数据
P_test = Test(:,3:end)';
T_test = Test(:,2)';
% 显示实验条件
total_B = length(find(data(:,2) == 1));
total_M = length(find(data(:,2) == 2));
count_B = length(find(T_train == 1));
count_M = length(find(T_train == 2));
number_B = length(find(T_test == 1));
number_M = length(find(T_test == 2));
disp('实验条件为：');
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)]);
%% 数据归一化
[P_train,minp,maxp,T_train,mint,maxt] = premnmx(P_train,T_train);
P_test = tramnmx(P_test,minp,maxp);
%% 创建单BP网络
t = cputime;
net_bp = newff(minmax(P_train),[s1,1],{'tansig','purelin'},'trainlm');
% 设置训练参数
net_bp.trainParam.epochs = 1000;
net_bp.trainParam.show = 10;
net_bp.trainParam.goal = 0.1;
net_bp.trainParam.lr = 0.1;
net_bp.trainParam.showwindow = 0;
%% 训练单BP网络
net_bp = train(net_bp,P_train,T_train);
%% 仿真测试单BP网络
tn_bp_sim = sim(net_bp,P_test);
% 反归一化
T_bp_sim = postmnmx(tn_bp_sim,mint,maxt);
e = cputime - t;
T_bp_sim(T_bp_sim > 1.5) = 2;
T_bp_sim(T_bp_sim < 1.5) = 1;
result_bp = [T_bp_sim' T_test'];
%% 结果显示（单BP网络）
number_B_sim = length(find(T_bp_sim == 1 & T_test == 1));
number_M_sim = length(find(T_bp_sim == 2 &T_test == 2));
disp('(1)BP网络的测试结果为：');
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) '%']);
disp(['建模时间为：' num2str(e) 's'] );
%% 遗传算法优化
popu = 20;  
bounds = ones(S,1)*[0,1];
% 产生初始种群
initPop = randint(popu,S,[0 1]);
% 计算初始种群适应度
initFit = zeros(popu,1);
for i = 1:size(initPop,1)
    initFit(i) = de_code(initPop(i,:));
end
initPop = [initPop initFit];
gen = 100; 
% 优化计算
[X,EndPop,BPop,Trace] = ga(bounds,'fitness',[],initPop,[1e-6 1 0],'maxGenTerm',...
    gen,'normGeomSelect',0.09,'simpleXover',2,'boundaryMutation',[2 gen 3]);
[m,n] = find(X == 1);
disp(['优化筛选后的输入自变量编号为:' num2str(n)]);
% 绘制适应度函数进化曲线
figure
plot(Trace(:,1),Trace(:,3),'r:')
hold on
plot(Trace(:,1),Trace(:,2),'b')
xlabel('进化代数')
ylabel('适应度函数')
title('适应度函数进化曲线')
legend('平均适应度函数','最佳适应度函数')
xlim([1 gen])
%% 新训练集/测试集数据提取
p_train = zeros(size(n,2),size(T_train,2));
p_test = zeros(size(n,2),size(T_test,2));
for i = 1:length(n)
    p_train(i,:) = P_train(n(i),:);
    p_test(i,:) = P_test(n(i),:);
end
t_train = T_train;
%% 创建优化BP网络
t = cputime;
net_ga = newff(minmax(p_train),[s1,1],{'tansig','purelin'},'trainlm');
% 训练参数设置
net_ga.trainParam.epochs = 1000;
net_ga.trainParam.show = 10;
net_ga.trainParam.goal = 0.1;
net_ga.trainParam.lr = 0.1;
net_ga.trainParam.showwindow = 0;
%% 训练优化BP网络
net_ga = train(net_ga,p_train,t_train);
%% 仿真测试优化BP网络
tn_ga_sim = sim(net_ga,p_test);
% 反归一化
T_ga_sim = postmnmx(tn_ga_sim,mint,maxt);
e = cputime - t;
T_ga_sim(T_ga_sim > 1.5) = 2;
T_ga_sim(T_ga_sim < 1.5) = 1;
result_ga = [T_ga_sim' T_test'];
%% 结果显示（优化BP网络）
number_b_sim = length(find(T_ga_sim == 1 & T_test == 1));
number_m_sim = length(find(T_ga_sim == 2 &T_test == 2));
disp('(2)优化BP网络的测试结果为：');
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) '%']);
disp(['建模时间为：' num2str(e) 's'] );

fitness

function [sol,Val] = fitness(sol,options)
global S
for i = 1:S
    x(i) = sol(i);
end
Val = de_code(x);
end

de_code

function Val = de_code(x)
% 全局变量声明
global S P_train T_train P_test T_test mint maxt 
global p t r s s1 s2
% 数据提取
x = x(:,1:S);
[m,n] = find(x == 1);
p_train = zeros(size(n,2),size(T_train,2));
p_test = zeros(size(n,2),size(T_test,2));
for i = 1:length(n)
    p_train(i,:) = P_train(n(i),:);
    p_test(i,:) = P_test(n(i),:);
end
t_train = T_train;
p = p_train;
t = t_train;
% 遗传算法优化BP网络权值和阈值
r = size(p,1);
s2 = size(t,1);
s = r*s1 + s1*s2 + s1 + s2;
aa = ones(s,1)*[-1,1];
popu = 20;  % 种群规模
initPpp = initializega(popu,aa,'gabpEval');  % 初始化种群
gen = 100;  % 遗传代数
% 调用GAOT工具箱，其中目标函数定义为gabpEval
x = ga(aa,'gabpEval',[],initPpp,[1e-6 1 0],'maxGenTerm',gen,...
'normGeomSelect',0.09,'arithXover',2,'nonUnifMutation',[2 gen 3]);
% 创建BP网络
net = newff(minmax(p_train),[s1,1],{'tansig','purelin'},'trainlm');
% 将优化得到的权值和阈值赋值给BP网络
[W1,B1,W2,B2] = gadecod(x);
net.IW{1,1} = W1;
net.LW{2,1} = W2;
net.b{1} = B1;
net.b{2} = B2;
% 设置训练参数
net.trainParam.epochs = 1000;
net.trainParam.show = 10;
net.trainParam.goal = 0.1;
net.trainParam.lr = 0.1;
net.trainParam.showwindow = 0;
% 训练网络
net = train(net,p_train,t_train);
% 仿真测试
tn_sim = sim(net,p_test);
% 反归一化
t_sim = postmnmx(tn_sim,mint,maxt);
% 计算均方误差
SE = sse(t_sim - T_test);
% 计算适应度函数值
Val = 1/SE;
end

gabpEval

function[sol,val] = gabpEval(sol,options)
global s
for i = 1:s
x(i) = sol(i);
end;
[W1,B1,W2,B2,val] = gadecod(x);

gadecod

function[W1,B1,W2,B2,val] = gadecod(x)
global p t r s1 s2
W1 = zeros(s1,r);
W2 = zeros(s2,s1);
B1 = zeros(s1,1);
B2 = zeros(s2,1);
% 前r*s1个编码为W1
for i = 1:s1
    for k = 1:r
        W1(i,k) = x(r*(i-1)+k);
    end
end
% 接着的s1*s2个编码(即第r*s1个后的编码)为W2
for i = 1:s2
    for k = 1:s1
        W2(i,k) = x(s1*(i-1)+k+r*s1);
    end
end
% 接着的s1个编码(即第r*s1+s1*s2个后的编码)为B1
for i = 1:s1
    B1(i,1) = x((r*s1+s1*s2)+i);
end
% 接着的s2个编码(即第r*s1+s1*s2+s1个后的编码)为B2
for i = 1:s2
    B2(i,1) = x((r*s1+s1*s2+s1)+i);
end
% 计算S1与S2层的输出
A1 = tansig(W1*p,B1);
A2 = purelin(W2*A1,B2);
% 计算误差平方和
SE = sumsqr(t - A2);
% 遗传算法的适应值
val = 1/SE;

如果需要数据文件可以后台私信我。