matlab faruto,faruto_psoSVMcgForRegress – MATLAB中文论坛

痛定思痛。 2023-01-15 14:21 148阅读 0赞

function \[bestCVmse,bestc,bestg,pso\_option\] = psoSVMcgForRegress(train\_label,train,pso\_option)

% psoSVMcgForClass

% by faruto

% 2009.10.28

%% 参数初始化

if nargin == 2

pso\_option = struct('c1',1.2,'c2',1,'maxgen',200,'sizepop',30, ...

'k',0.6,'wV',1,'wP',1,'v',5, ...

'popcmax',10^2,'popcmin',10^(-1),'popgmax',10^3,'popgmin',10^(-2));

end

% c1:初始为1.5,pso参数局部搜索能力

% c2:初始为1.7,pso参数全局搜索能力

% maxgen:初始为200,最大进化数量

% sizepop:初始为20,种群最大数量

% k:初始为0.6(k belongs to \[0.1,1.0\]),速率和x的关系(V = kX)

% wV:初始为1(wV best belongs to \[0.8,1.2\]),速率更新公式中速度前面的弹性系数

% wP:初始为1,种群更新公式中速度前面的弹性系数

% v:初始为3,SVM Cross Validation参数

% popcmax:初始为100,SVM 参数c的变化的最大值.

% popcmin:初始为0.1,SVM 参数c的变化的最小值.

% popgmax:初始为1000,SVM 参数g的变化的最大值.

% popgmin:初始为0.01,SVM 参数c的变化的最小值.

Vcmax = pso\_option.k\*pso\_option.popcmax;

Vcmin = -Vcmax ;

Vgmax = pso\_option.k\*pso\_option.popgmax;

Vgmin = -Vgmax ;

%% 产生初始粒子和速度

for i=1:pso\_option.sizepop

% 随机产生种群和速度

pop(i,1) = (pso\_option.popcmax-pso\_option.popcmin)\*rand+pso\_option.popcmin;

pop(i,2) = (pso\_option.popgmax-pso\_option.popgmin)\*rand+pso\_option.popgmin;

V(i,1)=Vcmax\*rands(1);

V(i,2)=Vgmax\*rands(1);

% 计算初始适应度

cmd = \['-v ',num2str(pso\_option.v),' -c ',num2str( pop(i,1) ),' -g ',num2str( pop(i,2) ),' -s 3 -p 0.1'\];

fitness(i) = svmtrain(train\_label, train, cmd);

end

% 找极值和极值点

\[global\_fitness bestindex\]=min(fitness); % 全局极值

local\_fitness=fitness;   % 个体极值初始化

global\_x=pop(bestindex,:);   % 全局极值点

local\_x=pop;    % 个体极值点初始化

%% 迭代寻优

for i=1:pso\_option.maxgen

for j=1:pso\_option.sizepop

%速度更新

V(j,:) = pso\_option.wV\*V(j,:) + pso\_option.c1\*rand\*(local\_x(j,:) - pop(j,:)) + pso\_option.c2\*rand\*(global\_x - pop(j,:));

if V(j,1) > Vcmax

V(j,1) = Vcmax;

end

if V(j,1) < Vcmin

V(j,1) = Vcmin;

end

if V(j,2) > Vgmax

V(j,2) = Vgmax;

end

if V(j,2) < Vgmin

V(j,2) = Vgmin;

end

%种群更新

pop(j,:)=pop(j,:) + pso\_option.wP\*V(j,:);

if pop(j,1) > pso\_option.popcmax

pop(j,1) = pso\_option.popcmax;

end

if pop(j,1) < pso\_option.popcmin

pop(j,1) = pso\_option.popcmin;

end

if pop(j,2) > pso\_option.popgmax

pop(j,2) = pso\_option.popgmax;

end

if pop(j,2) < pso\_option.popgmin

pop(j,2) = pso\_option.popgmin;

end

% 自适应粒子变异

if rand>0.5

k=ceil(2\*rand);

if k == 1

pop(j,k) = (20-1)\*rand+1;

end

if k == 2

pop(j,k) = (pso\_option.popgmax-pso\_option.popgmin)\*rand + pso\_option.popgmin;

end

%适应度值

cmd = \['-v ',num2str(pso\_option.v),' -c ',num2str( pop(j,1) ),' -g ',num2str( pop(j,2) ),' -s 3 -p 0.1'\];

fitness(j) = svmtrain(train\_label, train, cmd);

end

%个体最优更新

if fitness(j) < local\_fitness(j)

local\_x(j,:) = pop(j,:);

local\_fitness(j) = fitness(j);

end

%群体最优更新

if fitness(j) < global\_fitness

global\_x = pop(j,:);

global\_fitness = fitness(j);

end

fit\_gen(i)=global\_fitness;

end

%% 结果分析

figure;

plot(fit\_gen);

title(\['适应度曲线','(参数c1=',num2str(pso\_option.c1),',c2=',num2str(pso\_option.c2),',终止代数=',num2str(pso\_option.maxgen),',种群数量pop=',num2str(pso\_option.sizepop),')'\]);

xlabel('进化代数');ylabel('适应度');

bestc = global\_x(1);

bestg = global\_x(2);

bestCVmse = fit\_gen(pso\_option.maxgen);

k.jpg

(80.23 KB, 下载次数: 0)

2017-3-28 15:01 上传

![442a53943febe9465fc072b4fbe10813.gif][]

![b2a5a3e0dcc7d508e00275fe42fce1b5.gif][]

参数k何用？用它有何优点

![c3881289035060605699d256f4c71838.png][]

vgmax.png

(22.53 KB, 下载次数: 0)

2017-3-28 15:04 上传

![442a53943febe9465fc072b4fbe10813.gif][]

![b2a5a3e0dcc7d508e00275fe42fce1b5.gif][]

不一致啊，更新种群时用vc，更新速度为何不用vg呢？

![1a05299ae80464b8e31e1d10a4a98071.png][]

[442a53943febe9465fc072b4fbe10813.gif]: /images/20221022/396a180acb00471eba9971bc8b602170.png
[b2a5a3e0dcc7d508e00275fe42fce1b5.gif]: /images/20221022/e785468de2494f2eadaec704f97bb6fc.png
[c3881289035060605699d256f4c71838.png]: /images/20221022/278eca05247f49b6800f22aeb722025f.png
[1a05299ae80464b8e31e1d10a4a98071.png]: /images/20221022/5b74db9426784e309c4020e9edede915.png