【预测模型-ELM分类】基于鲸鱼算法优化核极限学习机实现数据分类附matlab代码

article/2025/3/5 10:42:08

1 内容介绍

极限学习机（ｅｘｔｒｅｍｅｌｅａｒｎｉｎｇｍａｃｈｉｎｅ，ＥＬＭ）作为一种新兴的机器学习方法，已经成为了一个热门的研究方向。ＥＬＭ随机确定单隐含层网络的输入权值和隐含层节点偏置，通过简单计算即可解析得到输出权值，不仅加快了学习速度，而且具有很好的泛化能力。核函数ＥＬＭ将核函数引入到极限学习机理论中，可以得到最小平方优化解，解决了

ＥＬＭ算法随机初始化问题，对模型学习参数具有较好的鲁棒性，具有更稳定的、更好的泛化性能。

2 仿真代码

function omega = kernel_matrix(Xtrain,kernel_type, kernel_pars,Xt)

nb_data = size(Xtrain,1);%nxm 样本数nb if strcmp(kernel_type,'RBF_kernel'), if nargin<4, XXh = sum(Xtrain.^2,2)*ones(1,nb_data);%nxn omega = XXh+XXh'-2*(Xtrain*Xtrain'); omega = exp(-omega./kernel_pars(1)); else XXh1 = sum(Xtrain.^2,2)*ones(1,size(Xt,1)); XXh2 = sum(Xt.^2,2)*ones(1,nb_data); omega = XXh1+XXh2' - 2*Xtrain*Xt'; omega = exp(-omega./kernel_pars(1)); end elseif strcmp(kernel_type,'lin_kernel') if nargin<4, omega = Xtrain*Xtrain'; else omega = Xtrain*Xt'; end elseif strcmp(kernel_type,'poly_kernel') if nargin<4, omega = (Xtrain*Xtrain'+kernel_pars(1)).^kernel_pars(2); else omega = (Xtrain*Xt'+kernel_pars(1)).^kernel_pars(2); end elseif strcmp(kernel_type,'wav_kernel') if nargin<4, XXh = sum(Xtrain.^2,2)*ones(1,nb_data); omega = XXh+XXh'-2*(Xtrain*Xtrain'); XXh1 = sum(Xtrain,2)*ones(1,nb_data); omega1 = XXh1-XXh1'; omega = cos(kernel_pars(3)*omega1./kernel_pars(2)).*exp(-omega./kernel_pars(1)); else XXh1 = sum(Xtrain.^2,2)*ones(1,size(Xt,1)); XXh2 = sum(Xt.^2,2)*ones(1,nb_data); omega = XXh1+XXh2' - 2*(Xtrain*Xt'); XXh11 = sum(Xtrain,2)*ones(1,size(Xt,1)); XXh22 = sum(Xt,2)*ones(1,nb_data); omega1 = XXh11-XXh22';

clc;clear;close all %% 将鲸鱼与改进鲸鱼用于函数极小值寻优，验证改进算法的有效性 % 优化函数极小值 popsize=10; Max_iter=100; dim=5; [best0,pop0,trace0]=woa(popsize,Max_iter,dim); best0 [best,pop,trace]=vnwoa(popsize,Max_iter,dim); best %% figure subplot(121) plot(trace0) hold on plot(trace) legend('woa','vnwoa')

subplot(122) plot(log10(trace0))%对结果取对数使结果更加直观 hold on plot(log10(trace)) legend('woa','vnwoa')

function y=boundary(x,hmin,hmax); [m n]=size(x); for i=1:m for j=1:n if x(i,j)>=hmax x(i,j)=rand*(hmax-hmin)+hmin; elseif x(i,j)<=hmin x(i,j)=rand*(hmax-hmin)+hmin; end end y=x;

end%

% ylabel('样本标签') % legend('真实标签', '预测标签'); % title(['KELM分类精度为：',num2str(acc*100),'%']) xlabel('Number of test samples'); ylabel('Class label'); legend('Actual output label','Desired output label'); title(['Accuracy：',num2str(acc*100),'%']); legend('boxoff'); hold off; omega = cos(kernel_pars(3)*omega1./kernel_pars(2)).*exp(-omega./kernel_pars(1)); end end

function [Leader_score,Leader_pos,Convergence_curve]=woa(sizepop,Max_iter,dim) lb=-100; ub=100; Convergence_curve=zeros(1,Max_iter); for i=1:sizepop Positions(i,:)=rand(1,dim).*(ub-lb)+lb; end for i=1:sizepop p(i)=fun(Positions(i,:)); %计算当前个体适应度值 end [~, index]=min(p); Leader_pos=Positions(index,:); Leader_score=p(index);

% Main loop for t=1:Max_iter a=5-t*((2)/Max_iter); a2=-1+t*((-1)/Max_iter); for i=1:size(Positions,1) r1=rand; r2=rand; A=2*a*r1-a; C=2*r2; b=1; l=(a2-1)*rand+1; p = rand; for j=1:size(Positions,2) if p<0.5 if abs(A)>=1 rand_leader_index = floor(sizepop*rand+1); X_rand = Positions(rand_leader_index, :); D_X_rand=abs(C*X_rand(j)-Positions(i,j)); Positions(i,j)=X_rand(j)-A*D_X_rand; elseif abs(A)<1 D_Leader=abs(C*Leader_pos(j)-Positions(i,j)); Positions(i,j)=Leader_pos(j)-A*D_Leader; end elseif p>=0.5 distance2Leader=abs(Leader_pos(j)-Positions(i,j)); Positions(i,j)=distance2Leader*exp(b.*l).*cos(l.*2*pi)+Leader_pos(j); end end Positions(i,:)=boundary(Positions(i,:),lb,ub); fitness=fun(Positions(i,:)); if fitness<Leader_score Leader_score=fitness; Leader_pos=Positions(i,:); end end Convergence_curve(t)=Leader_score; end

function f =luoyiman_localbest(x1,x2,x3,x4,x5)
%求解当前个体粒子环形邻域中的局部最优,然后用这个局部最优替换当前个体
K0=[x1;x2;x3;x4;x5];
K1=[fun(x1), ...
fun(x2), ...
fun(x3), ...
fun(x4), ...
fun(x5)];
[~ ,index]=min(K1);
f=K0(index,:);

function fit=fun(pop) % 这3种函数的最小值都是0

%1.sphere函数 % fit=sum(pop.^2);

%2. ackly函数 % fit=-20*exp(-0.2*sqrt(mean(pop.^2)))-exp(mean(cos(2*pi.*pop)))+20+exp(1);

% 3.rastrigin's函数 fit=20+pop(1)^2+pop(2)^2-10*(cos(2*pi*pop(1))+cos(2*pi*pop(2)));