Perceptron

perceptron.cpp

#include <iostream>
#include <vector>
#include <numeric>

using namespace std;

class Perceptron {
    vector<double> w;
    double learning_rate;
    bool error;

public:

    Perceptron(vector<double> _w, double _learning_rate) {
        w = _w;
        learning_rate = _learning_rate;
        error = false;
    };

    ~Perceptron() {};

    vector<double> pred(vector<double> x);
    void train(vector<vector<double>> x, vector<double> y);
    double sum(vector<double> xw);
};

vector<double> Perceptron::pred(vector<double> x) {
    vector<double>::const_iterator x_it = x.begin();
    vector<double>::const_iterator w_it = w.begin();

    vector<double> y = vector<double>();

    for (;x_it != x.end() && w_it != w.end(); ++x_it, ++w_it) {
        y.push_back((*x_it) * (*w_it));
    }
    
    return y;
}

double add(double x1, double x2) {
    return x1 + x2;
}

int sign(double x) {
    return (x > 0) ? 1 : ((x < 0) ? -1 : 0);
}

double Perceptron::sum(vector<double> xw) {
    double y = 0.0;
    for (auto _y : xw) {
        y += _y;
    }
    return y;
}

void Perceptron::train(vector<vector<double>> xs, vector<double> ys) {
    vector<vector<double>>::const_iterator xs_it = xs.begin();
    vector<double>::const_iterator ys_it = ys.begin();
    int c = 0;

    error = true;
    while (error) {
        c++;
        error = false;
        for (;xs_it != xs.end() && ys_it != ys.end(); ++xs_it, ++ys_it) {
            double y_hat = sum(pred(*xs_it));
            double y = *ys_it * y_hat;
            if (sign(y) <= 0) {
                for (size_t i = 0; i < w.size(); i ++) {
                    w[i] += y * (*xs_it)[i];
                }
                error = true;
            }
        }
    }

    cout << "Iterations to convergence: " << c << endl;
    error = false;
}

int main(int argc, char *argv[]) {
    vector<double> x({0.01, 0.01});
    Perceptron p(x, 0.02);

    vector<double> x1({0.0, 0.0});
    vector<double> x2({1.0, 0.0});
    vector<double> x3({0.0, 1.0});
    vector<double> x4({1.0, 1.0});

    double y1 = 0.0;
    double y2 = 1.0;
    double y3 = 1.0;
    double y4 = 1.0;

    vector<vector<double>> dataset({x1, x2, x3, x4});
    vector<double> ys({y1, y2, y3, y4});

    p.train(dataset, ys);

}