fork download
copy 
  1. #include <iostream>
  2. #include <vector>
  3.  
  4. // Tu wklejam klasę Perceptron z poprzedniego przykładu
  5.  
  6. class Perceptron {
  7. private:
  8.     std::vector<double> weights;
  9.     double bias;
  10.     double learning_rate;
  11.  
  12. public:
  13.     Perceptron(int input_size, double learning_rate)
  14.         : weights(input_size, 0.0), bias(0.0), learning_rate(learning_rate) {}
  15.  
  16.     int predict(const std::vector<double>& inputs) {
  17.         double suma = bias;
  18.         for (size_t i = 0; i < weights.size(); ++i) {
  19.             suma += weights[i] * inputs[i];
  20.         }
  21.         return (suma >= 0) ? 1 : 0;
  22.     }
  23.  
  24.     void train(const std::vector<std::vector<double>>& inputs,
  25.                const std::vector<int>& labels,
  26.                int epochs) {
  27.         for (int epoch = 0; epoch < epochs; ++epoch) {
  28.             for (size_t i = 0; i < inputs.size(); ++i) {
  29.                 int prediction = predict(inputs[i]);
  30.                 int error = labels[i] - prediction;
  31.  
  32.                 for (size_t j = 0; j < weights.size(); ++j) {
  33.                     weights[j] += learning_rate * error * inputs[i][j];
  34.                 }
  35.                 bias += learning_rate * error;
  36.             }
  37.         }
  38.     }
  39.  
  40.     void printWeightsAndBias() const {
  41.         std::cout << "Wagi: ";
  42.         for (double w : weights) {
  43.             std::cout << w << " ";
  44.         }
  45.         std::cout << "\nBias: " << bias << std::endl;
  46.     }
  47. };
  48.  
  49. int main() {
  50.     // Tworzymy perceptron z 2 wejściami i współczynnikiem uczenia 0.1
  51.     Perceptron p(2, 0.1);
  52.  
  53.     // Dane treningowe (wejścia)
  54.     std::vector<std::vector<double>> training_inputs = {
  55.         {0, 0},
  56.         {0, 1},
  57.         {1, 0},
  58.         {1, 1}
  59.     };
  60.  
  61.     // Etykiety zgodne z funkcją AND, ale wyjścia to 0 lub 2 (zgodnie z Twoim kodem)
  62.     std::vector<int> training_labels = {0, 0, 0, 1};
  63.  
  64.     // Trenujemy perceptron przez 10 epok
  65.     p.train(training_inputs, training_labels, 10);
  66.  
  67.     // Wypisujemy wagi i bias po treningu
  68.     p.printWeightsAndBias();
  69.  
  70.     // Testujemy perceptron na danych treningowych
  71.     std::cout << "Test predykcji:" << std::endl;
  72.     for (const auto& input : training_inputs) {
  73.         int result = p.predict(input);
  74.         std::cout << "(" << input[0] << ", " << input[1] << ") -> " << result << std::endl;
  75.     }
  76.  
  77.     return 0;
  78. }
  79.  
Success #stdin #stdout 0.01s 5288KB
comments (?)
stdin
copy 
Standard input is empty
stdout
copy 
Wagi: 0.2 0.1 
Bias: -0.2
Test predykcji:
(0, 0) -> 0
(0, 1) -> 0
(1, 0) -> 0
(1, 1) -> 1
https://ideone.com/PAKtlU
language:
C++ (gcc 8.3)
created:
1 day ago
visibility:
public

Share or Embed source code

Discover > Sphere Engine API

The brand new service which powers Ideone!

Discover > IDE Widget

Widget for compiling and running the source code in a web browser!