VMD-GRU/GRU基于变分模态分解和门控循环单元的时间序列预测(含GRU、VMD-GRU模型的对比)。评价指标包括:R2、MAE、MSE、RMSE和MAPE等，代码质量极高，方便学习和替换数据

 

%% 建立
T_sim3 =[];
T_sim4 =[];
data1 = u';

for i = 1:size(data1,2)
data2=  data1(:,i);
num_samples = length(data2);       % 样本个数 
kim = 5;                      % 延时步长（kim个历史数据作为自变量）
zim =  1;                      % 跨zim个时间点进行预测
or_dim = size(data2,2);
res=[];
%  重构数据集
for i = 1: num_samples - kim - zim + 1
    res(i, :) = [reshape(data2(i: i + kim - 1,:), 1, kim*or_dim), data2(i + kim + zim - 1,:)];
end

% 训练集和测试集划分
outdim = 1;                                  % 最后一列为输出
num_size = 0.7;                              % 训练集占数据集比例
num_train_s = round(num_size * num_samples); % 训练集样本个数
f_ = size(res, 2) - outdim;                  % 输入特征维度

P_train = res(1: num_train_s, 1: f_)';
T_train = res(1: num_train_s, f_ + 1: end)';
M = size(P_train, 2);

P_test = res(num_train_s + 1: end, 1: f_)';
T_test = res(num_train_s + 1: end, f_ + 1: end)';
N = size(P_test, 2);

%  数据归一化
[p_train, ps_input] = mapminmax(P_train, 0, 1);
p_test = mapminmax('apply', P_test, ps_input);

[t_train, ps_output] = mapminmax(T_train, 0, 1);
t_test = mapminmax('apply', T_test, ps_output);

%  格式转换
for i = 1 : M 
    vp_train{i, 1} = p_train(:, i);
    vt_train{i, 1} = t_train(:, i);
end

for i = 1 : N 
    vp_test{i, 1} = p_test(:, i);
    vt_test{i, 1} = t_test(:, i);
end
 

智能算法及其模型预测