回歸預(yù)測 | MATLAB實現(xiàn)WOA-DNN鯨魚算法優(yōu)化深度神經(jīng)網(wǎng)絡(luò)的數(shù)據(jù)多輸入單輸出回歸預(yù)測

效果一覽

基本介紹

回歸預(yù)測 | MATLAB實現(xiàn)WOA-DNN鯨魚算法優(yōu)化深度神經(jīng)網(wǎng)絡(luò)的數(shù)據(jù)多輸入單輸出回歸預(yù)測
MATLAB實現(xiàn)WOA-DNN鯨魚算法優(yōu)化深度神經(jīng)網(wǎng)絡(luò)的數(shù)據(jù)多輸入單輸出回歸預(yù)測（Matlab完整程序和數(shù)據(jù)）
輸入7個特征，輸出1個，即多輸入單輸出；優(yōu)化參數(shù)為學(xué)習(xí)率，批大小，正則化系數(shù)。
運行環(huán)境Matlab2018及以上,運行主程序main即可，其余為函數(shù)文件無需運行,所有程序放在一個文件夾,data為數(shù)據(jù)集；
命令窗口輸出RMSE、MAE、R2、MAPE。

程序設(shè)計

• 完整程序和數(shù)據(jù)下載方式1(訂閱《DNN深度神經(jīng)網(wǎng)絡(luò)》專欄，同時可閱讀《DNN深度神經(jīng)網(wǎng)絡(luò)》專欄收錄的所有內(nèi)容，數(shù)據(jù)訂閱后私信我獲取)：MATLAB實現(xiàn)WOA-DNN鯨魚算法優(yōu)化深度神經(jīng)網(wǎng)絡(luò)的數(shù)據(jù)多輸入單輸出回歸預(yù)測
• 完整程序和數(shù)據(jù)下載方式2(訂閱《智能學(xué)習(xí)》專欄，同時獲取《智能學(xué)習(xí)》專欄收錄程序3份，數(shù)據(jù)訂閱后私信我獲取)：MATLAB實現(xiàn)WOA-DNN鯨魚算法優(yōu)化深度神經(jīng)網(wǎng)絡(luò)的數(shù)據(jù)多輸入單輸出回歸預(yù)測

%%  記錄最佳參數(shù)
Best_pos(1, 2) = round(Best_pos(1, 2));
best_lr = Best_pos(1, 1);
best_hd = Best_pos(1, 2);
best_l2 = Best_pos(1, 3);

%%  建立模型
% ----------------------  修改模型結(jié)構(gòu)時需對應(yīng)修改fical.m中的模型結(jié)構(gòu)  --------------------------
layers = [
    sequenceInputLayer(f_)            % 輸入層
   
    fullyConnectedLayer(outdim)       % 輸出回歸層
    regressionLayer];
 
%%  參數(shù)設(shè)置
% ----------------------  修改模型參數(shù)時需對應(yīng)修改fical.m中的模型參數(shù)  --------------------------
options = trainingOptions('adam', ...           % Adam 梯度下降算法
         'MaxEpochs', 500, ...                  % 最大訓(xùn)練次數(shù) 500
         'InitialLearnRate', best_lr, ...       % 初始學(xué)習(xí)率 best_lr
         'LearnRateSchedule', 'piecewise', ...  % 學(xué)習(xí)率下降
         'LearnRateDropFactor', 0.5, ...        % 學(xué)習(xí)率下降因子 0.1
         'LearnRateDropPeriod', 400, ...        % 經(jīng)過 400 次訓(xùn)練后 學(xué)習(xí)率為 best_lr * 0.5
         'Shuffle', 'every-epoch', ...          % 每次訓(xùn)練打亂數(shù)據(jù)集
         'ValidationPatience', Inf, ...         % 關(guān)閉驗證
         'L2Regularization', best_l2, ...       % 正則化參數(shù)
         'Plots', 'training-progress', ...      % 畫出曲線
         'Verbose', false);

%%  訓(xùn)練模型
net = trainNetwork(p_train, t_train, layers, options);

%%  仿真驗證
t_sim1 = predict(net, p_train);
t_sim2 = predict(net, p_test );

%%  數(shù)據(jù)反歸一化
T_sim1 = mapminmax('reverse', t_sim1, ps_output);
T_sim2 = mapminmax('reverse', t_sim2, ps_output);
T_sim1=double(T_sim1);
T_sim2=double(T_sim2);
%%  均方根誤差
error1 = sqrt(sum((T_sim1 - T_train).^2) ./ M);
error2 = sqrt(sum((T_sim2 - T_test ).^2) ./ N);
%_________________________________________________________________________%
% The Whale Optimization Algorithm
function [Best_Cost,Best_pos,curve]=WOA(pop,Max_iter,lb,ub,dim,fobj)

% initialize position vector and score for the leader
Best_pos=zeros(1,dim);
Best_Cost=inf; %change this to -inf for maximization problems


%Initialize the positions of search agents
Positions=initialization(pop,dim,ub,lb);

curve=zeros(1,Max_iter);

t=0;% Loop counter

% Main loop
while t<Max_iter
    for i=1:size(Positions,1)
        
        % Return back the search agents that go beyond the boundaries of the search space
        Flag4ub=Positions(i,:)>ub;
        Flag4lb=Positions(i,:)<lb;
        Positions(i,:)=(Positions(i,:).*(~(Flag4ub+Flag4lb)))+ub.*Flag4ub+lb.*Flag4lb;
        
        % Calculate objective function for each search agent
        fitness=fobj(Positions(i,:));
        
        % Update the leader
        if fitness<Best_Cost % Change this to > for maximization problem
            Best_Cost=fitness; % Update alpha
            Best_pos=Positions(i,:);
        end
        
    end
    
    a=2-t*((2)/Max_iter); % a decreases linearly fron 2 to 0 in Eq. (2.3)
    
    % a2 linearly dicreases from -1 to -2 to calculate t in Eq. (3.12)
    a2=-1+t*((-1)/Max_iter);
    
    % Update the Position of search agents 
    for i=1:size(Positions,1)
        r1=rand(); % r1 is a random number in [0,1]
        r2=rand(); % r2 is a random number in [0,1]
        
        A=2*a*r1-a;  % Eq. (2.3) in the paper
        C=2*r2;      % Eq. (2.4) in the paper
        
        
        b=1;               %  parameters in Eq. (2.5)
        l=(a2-1)*rand+1;   %  parameters in Eq. (2.5)
        
        p = rand();        % p in Eq. (2.6)
        
        for j=1:size(Positions,2)
            
            if p<0.5   
                if abs(A)>=1
                    rand_leader_index = floor(pop*rand()+1);
                    X_rand = Positions(rand_leader_index, :);
                    D_X_rand=abs(C*X_rand(j)-Positions(i,j)); % Eq. (2.7)
                    Positions(i,j)=X_rand(j)-A*D_X_rand;      % Eq. (2.8)
                    
                elseif abs(A)<1
                    D_Leader=abs(C*Best_pos(j)-Positions(i,j)); % Eq. (2.1)
                    Positions(i,j)=Best_pos(j)-A*D_Leader;      % Eq. (2.2)
                end
                
            elseif p>=0.5
              
                distance2Leader=abs(Best_pos(j)-Positions(i,j));
                % Eq. (2.5)
                Positions(i,j)=distance2Leader*exp(b.*l).*cos(l.*2*pi)+Best_pos(j);
                
            end
            
        end
    end
    t=t+1;
    curve(t)=Best_Cost;
    [t Best_Cost]
end

參考資料

[1] https://blog.csdn.net/kjm13182345320/article/details/129215161
[2] https://blog.csdn.net/kjm13182345320/article/details/128105718文章來源地址http://www.zghlxwxcb.cn/news/detail-543721.html

到了這里，關(guān)于回歸預(yù)測 | MATLAB實現(xiàn)WOA-DNN鯨魚算法優(yōu)化深度神經(jīng)網(wǎng)絡(luò)的數(shù)據(jù)多輸入單輸出回歸預(yù)測的文章就介紹完了。如果您還想了解更多內(nèi)容，請在右上角搜索TOY模板網(wǎng)以前的文章或繼續(xù)瀏覽下面的相關(guān)文章，希望大家以后多多支持TOY模板網(wǎng)！