感知器算法 C++

We can estimate the weight values for our training data using stochastic gradient descent.

Stochastic gradient descent requires two parameters:

Learning Rate: Used to limit the amount each weight is corrected each time it is updated.
Epochs: The number of times to run through the training data while updating the weight.

These, along with the training data will be the arguments to the function.

There are 3 loops we need to perform in the function:

Loop over each epoch.
Loop over each row in the training data for an epoch.
Loop over each weight and update it for a row in an epoch.

As you can see, we update each weight for each row in the training data, each epoch.

The loop is over until:

　　the iteration error is less than a user-specified error threshold or

　　a predetermined number of iterations have been completed.

Weights are updated based on the error the model made. The error is calculated as the difference between the expected output value and the prediction made with the candidate weights.

Notice that learning only occurs when an error is made, otherwise the weights are left unchanged.

#include <iostream>
#include <string>
#include <fstream>
#include <sstream>
#include <vector>
#include <cmath>

//the sign function
template <typename DataType, typename WeightType>
double sign(typename::std::vector<DataType> &data, typename::std::vector<WeightType> &weights) {
double result=0.0;

for(size_t i=0; i<weights.size(); ++i) {
result += data.at(i)*weights.at(i);
}

if(result >= 0.0)
return 1.0;
else
return 0.0;
}

template <typename DataType, typename WeightType>
void trainW(typename::std::vector<std::vector<DataType> > &vv, typename::std::vector<WeightType> &weights, const double& l_rate, const int& n_epoch) {
std::vector<DataType> v_data;

for(size_t i=0; i<weights.size(); ++i) {
weights.at(i)=0.0;
}

for(size_t i=0; i<n_epoch; ++i) {
double sum_error=0.0;

for(size_t j=0; j<vv.size(); ++j) {
v_data.clear();
for(size_t k=0; k<weights.size(); ++k) {
v_data.push_back(vv[j][k]);
}

for(typename::std::vector<DataType>::iterator it=v_data.begin();it!=v_data.end();++it) {
std::cout<<*it<<" ";
}

std::cout<<std::endl;

double prediction=sign(v_data, weights);
double error=vv[j].back()-prediction;
std::cout<<"expected: "<<vv[j].back()<<" prediction: "<<prediction<<" error: "<<error<<std::endl;

sum_error+=pow(error, 2.0);

for(size_t k=0; k<weights.size(); ++k) {
weights.at(k)=weights.at(k)+l_rate*error*vv[j][k];
}
}
std::cout<<"epoch = "<<i<<" error = "<<sum_error<<std::endl;
}

for(size_t i=0; i<weights.size(); ++i) {
std::cout<<weights.at(i)<<" ";
}
std::cout<<std::endl;
}

//make a prediction with weights, appended to the last column
template <typename DataType, typename WeightType>
void predictTestData(typename::std::vector<std::vector<DataType> > &vv, typename::std::vector<WeightType> &weights) {
std::vector<DataType> v_data;

for(size_t i=0;i<vv.size();++i) {
v_data.clear();
for(size_t j=0;j<weights.size();++j) {
v_data.push_back(vv[i][j]);
}

double signResult=sign(v_data,weights);
vv[i].push_back(signResult);
}
}

//display the data
template <typename DataType>
void DisplayData(typename::std::vector<std::vector<DataType> > &vv) {
std::cout<<"the number of data: "<<vv.size()<<std::endl;

for(size_t i=0; i<vv.size(); ++i) {
for(typename::std::vector<DataType>::iterator it=vv[i].begin(); it!=vv[i].end(); ++it) {
std::cout<<*it<<" ";
}
std::cout<<std::endl;
}
}

int main() {
std::ifstream infile_feat("PLA.txt");
std::string feature;
float feat_onePoint;
std::vector<float> lines;
std::vector<std::vector<float> > lines_feat;
lines_feat.clear();

std::vector<float> v_weights;
v_weights.clear();
v_weights.push_back(-0.1);
v_weights.push_back(0.206);
v_weights.push_back(-0.234);

while(!infile_feat.eof()) {
getline(infile_feat, feature);
if(feature.empty())
break;
std::stringstream stringin(feature);
lines.clear();

lines.push_back(1);
while(stringin >> feat_onePoint) {
lines.push_back(feat_onePoint);
}
lines_feat.push_back(lines);
}

infile_feat.close();

std::cout<<"display train data: "<<std::endl;

DisplayData(lines_feat);

double l_rate=0.1;

int n_epoch=5;

trainW(lines_feat, v_weights, l_rate, n_epoch);

//predictTestData(lines_feat, v_weights);

//std::cout<<"the predicted: "<<std::endl;
//DisplayData(lines_feat);

return 0;
}

巴特西

感知器算法 C++

最新文章

热门文章