改用c++实现bp神经网络

以下是使用C++实现BP神经网络的步骤：

1.定义神经元结构体

struct Neuron {
    double output; // 神经元输出值
    double gradient; // 神经元梯度
    vector<double> outputWeights; // 神经元输出权重
    vector<double> deltaOutputWeights; // 神经元输出权重的变化量
    unsigned index; // 神经元索引
    double sumDOW(const Layer &nextLayer) const; // 计算神经元的误差项
    void calcOutputGradients(double targetVal); // 计算输出层神经元的梯度
    void calcHiddenGradients(const Layer &nextLayer); // 计算隐藏层神经元的梯度
    void updateInputWeights(Layer &prevLayer); // 更新神经元的输入权重
};

2.定义层结构体

typedef vector<Neuron> Layer;

class Net {
public:
    Net(const vector<unsigned> &topology); // 构造函数
    void feedForward(const vector<double> &inputVals); // 前向传播
    void backProp(const vector<double> &targetVals); // 反向传播
    void getResults(vector<double> &resultVals) const; // 获取输出层的结果
    double getRecentAverageError() const { return m_recentAverageError; } // 获取最近的平均误差
private:
    vector<Layer> m_layers; // 神经网络的层
    double m_error; // 神经网络的误差
    double m_recentAverageError; // 最近的平均误差
    double m_recentAverageSmoothingFactor; // 平均误差的平滑因子
};

3.实现构造函数

Net::Net(const vector<unsigned> &topology) {
    unsigned numLayers = topology.size();
    for (unsigned layerNum = 0; layerNum < numLayers; ++layerNum) {
        m_layers.push_back(Layer());
        unsigned numOutputs = layerNum == topology.size() - 1 ? 0 : topology[layerNum + 1];
        for (unsigned neuronNum = 0; neuronNum <= topology[layerNum]; ++neuronNum) {
            m_layers.back().push_back(Neuron(numOutputs, neuronNum));
            cout << "Made a Neuron!" << endl;
        }
        m_layers.back().back().setOutputVal(1.0);
    }
}

4.实现前向传播

void Net::feedForward(const vector<double> &inputVals) {
    assert(inputVals.size() == m_layers[0].size() - 1);
    for (unsigned i = 0; i < inputVals.size(); ++i) {
        m_layers[0][i].setOutputVal(inputVals[i]);
    }
    for (unsigned layerNum = 1; layerNum < m_layers.size(); ++layerNum) {
        Layer &prevLayer = m_layers[layerNum - 1];
        for (unsigned n = 0; n < m_layers[layerNum].size() - 1; ++n) {
            m_layers[layerNum][n].feedForward(prevLayer);
        }
    }
}

5.实现反向传播

void Net::backProp(const vector<double> &targetVals) {
    Layer &outputLayer = m_layers.back();
    m_error = 0.0;
    for (unsigned n = 0; n < outputLayer.size() - 1; ++n) {
        double delta = targetVals[n] - outputLayer[n].getOutputVal();
        m_error += delta * delta;
    }
    m_error /= outputLayer.size() - 1;
    m_error = sqrt(m_error);
    m_recentAverageError = (m_recentAverageError * m_recentAverageSmoothingFactor + m_error) / (m_recentAverageSmoothingFactor + 1.0);
    for (unsigned n = 0; n < outputLayer.size() - 1; ++n) {
        outputLayer[n].calcOutputGradients(targetVals[n]);
    }
    for (unsigned layerNum = m_layers.size() - 2; layerNum > 0; --layerNum) {
        Layer &hiddenLayer = m_layers[layerNum];
        Layer &nextLayer = m_layers[layerNum + 1];
        for (unsigned n = 0; n < hiddenLayer.size(); ++n) {
            hiddenLayer[n].calcHiddenGradients(nextLayer);
        }
    }
    for (unsigned layerNum = m_layers.size() - 1; layerNum > 0; --layerNum) {
        Layer &layer = m_layers[layerNum];
        Layer &prevLayer = m_layers[layerNum - 1];
        for (unsigned n = 0; n < layer.size() - 1; ++n) {
            layer[n].updateInputWeights(prevLayer);
        }
    }
}

6.实现获取输出层结果

void Net::getResults(vector<double> &resultVals) const {
    resultVals.clear();
    for (unsigned n = 0; n < m_layers.back().size() - 1; ++n) {
        resultVals.push_back(m_layers.back()[n].getOutputVal());
    }
}