clear

clc

% 準備數據

inputs = rand(10, 100); % 100組輸入，每組10個特征

outputs = rand(1, 100); % 100組輸出，每組1個輸出值

% 將數據分成訓練集和測試集

trainRatio = 0.8;

valRatio = 0.1;

testRatio = 0.1;

[trainInd, valInd, testInd] = dividerand(100, trainRatio, valRatio, testRatio);

X_train = inputs(:, trainInd);

y_train = outputs(:, trainInd);

X_test = inputs(:, testInd);

y_test = outputs(:, testInd);

% 創建前饋神經網絡

hiddenLayerSize = 10;

net = feedforwardnet(hiddenLayerSize);

% 設置訓練、驗證和測試數據

net.divideParam.trainRatio = 0.8;

net.divideParam.valRatio = 0.1;

net.divideParam.testRatio = 0.1;

% 訓練神經網絡

[net, tr] = train(net, X_train, y_train);

% 查看訓練結果

plotperform(tr);

% 進行預測

y_pred = net(X_test);

% 計算誤差

errors = y_pred - y_test;

% 輸出誤差

for i = 1:length(y_test)

fprintf('Test sample %d:\n', i);

fprintf('Predicted output: %.4f\n', y_pred(i));

fprintf('Actual output: %.4f\n', y_test(i));

fprintf('Error: %.4f\n\n', errors(i));

end

% 計算均方誤差（MSE）

mse = mean(errors.^2);

fprintf('Mean Squared Error: %.4f\n', mse);