4
\$\begingroup\$

I am a new to Eigen, and I implemented a logistic regression model with it. It works but I don't know whether it is implemented in an efficient way.

#include <iostream>
#include <Eigen/Dense>
#include <cmath>
using namespace Eigen;
using namespace Eigen::internal;
using namespace Eigen::Architecture;
using namespace std;
class logistic_regression
{
public:
 VectorXd w;
 double b;
 logistic_regression(int n_in)
 {
 this->w = VectorXd::Random(n_in);
 this->b = 0.0;
 }
 void train(MatrixXd train_datas, double lr)
 {
 VectorXd dw(this->w.rows());
 double db;
 dw = train_datas.rightCols(1) - calc(train_datas.leftCols(this->w.rows()));
 db = dw.mean();
 MatrixXd tmp = train_datas.leftCols(this->w.rows());
 tmp = tmp.array().colwise()*dw.array();
 dw = tmp.colwise().mean();
 w += lr * dw;
 b += lr * db;
 }
 VectorXd predict(MatrixXd inputs)
 {
 return ((1 / (1 + ((-inputs*this->w).array() - b).exp())).array() > 0.5).cast<double>();
 }
 double test_error(MatrixXd datas)
 {
 VectorXd outputs = predict(datas.leftCols(w.rows()));
 /*
 cout << outputs << endl << endl << endl;
 cout << datas.rightCols(1) << endl << endl << endl;
 cout << (outputs - datas.rightCols(1)) << endl;
 cout << ((outputs - datas.rightCols(1)).array() > 1e-5).count() << endl;*/
 return ((outputs - datas.rightCols(1)).array() > 1e-5).count() / (double)outputs.rows();
 }
};
MatrixXd linear_separable_dataset_generator(int n, int dim, VectorXd w, double b)
{
 int n_col = w.rows();
 MatrixXd datas = MatrixXd::Random(n, n_col + 1);
 datas.rightCols(1) = (((datas.leftCols(n_col)*w).array() + b) > 0).cast<double>();
 return datas;
}
void test_logistic_regression()
{
 logistic_regression lr(2);
 VectorXd w(2);
 double b = 0.2;
 w << 0.3, 0.6;
 MatrixXd train_datas = linear_separable_dataset_generator(1000, 2, w, b);
 for (int i = 0; i < 500; i++)
 {
 lr.train(train_datas.topRows(990), 0.1112);
 cout << "epoch:" << i << " error:" << lr.test_error(train_datas) << endl;
 }
 cout << "w:" << lr.w / (lr.b / 0.2) << endl;
 cout << "b:" << lr.b << endl;
 getchar();
 return;
}
200_success
145k22 gold badges190 silver badges478 bronze badges
asked Dec 3, 2015 at 13:44
\$\endgroup\$
0

0

Know someone who can answer? Share a link to this question via email, Twitter, or Facebook.

Your Answer

Draft saved
Draft discarded

Sign up or log in

Sign up using Google
Sign up using Email and Password

Post as a guest

Required, but never shown

Post as a guest

Required, but never shown

By clicking "Post Your Answer", you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.