matlab实现bp神经网络分类代码

以下是一个简单的 MATLAB BP 神经网络分类代码示例：

% 载入数据
data = load('data.csv');
X = data(:, 1:end-1);  % 特征
y = data(:, end);     % 目标变量

% 数据预处理
X = normalize(X);  % 特征标准化

% 分割数据集为训练集和测试集
[trainInd, valInd, testInd] = dividerand(size(X, 1), 0.6, 0.2, 0.2);
X_train = X(trainInd, :);
y_train = y(trainInd, :);
X_val = X(valInd, :);
y_val = y(valInd, :);
X_test = X(testInd, :);
y_test = y(testInd, :);

% 定义神经网络结构
input_layer_size = size(X, 2);
hidden_layer_size = 10;
output_layer_size = 1;

% 初始化神经网络权重
initial_Theta1 = randInitializeWeights(input_layer_size, hidden_layer_size);
initial_Theta2 = randInitializeWeights(hidden_layer_size, output_layer_size);
initial_nn_params = [initial_Theta1(:) ; initial_Theta2(:)];

% 训练神经网络
lambda = 1;
options = optimset('MaxIter', 100);
costFunction = @(p) nnCostFunction(p, input_layer_size, hidden_layer_size, output_layer_size, X_train, y_train, lambda);
[nn_params, cost] = fmincg(costFunction, initial_nn_params, options);

% 用训练好的神经网络进行预测
pred_train = predict(Theta1, Theta2, X_train);
pred_test = predict(Theta1, Theta2, X_test);

% 计算准确率
accuracy_train = mean(double(pred_train == y_train)) * 100;
accuracy_test = mean(double(pred_test == y_test)) * 100;

fprintf('Train Accuracy: %f\n', accuracy_train);
fprintf('Test Accuracy: %f\n', accuracy_test);

其中，`normalize` 函数用于标准化特征，`dividerand` 函数用于将数据集分割为训练集、验证集和测试集，`randInitializeWeights` 函数用于随机初始化权重，`nnCostFunction` 函数计算神经网络的代价函数和梯度，`fmincg` 函数用于优化代价函数，`predict` 函数用于进行预测，最后计算训练集和测试集的准确率。