Line data    Source code

       1             : /*************************************************************************
       2             :  *
       3             :  *  Project
       4             :  *                         _____ _____  __  __ _____
       5             :  *                        / ____|  __ \|  \/  |  __ \
       6             :  *  ___  _ __   ___ _ __ | |  __| |__) | \  / | |__) |
       7             :  * / _ \| '_ \ / _ \ '_ \| | |_ |  ___/| |\/| |  ___/
       8             :  *| (_) | |_) |  __/ | | | |__| | |    | |  | | |
       9             :  * \___/| .__/ \___|_| |_|\_____|_|    |_|  |_|_|
      10             :  *      | |
      11             :  *      |_|
      12             :  *
      13             :  * Copyright (C) Akiel Aries, <akiel@akiel.org>, et al.
      14             :  *
      15             :  * This software is licensed as described in the file LICENSE, which
      16             :  * you should have received as part of this distribution. The terms
      17             :  * among other details are referenced in the official documentation
      18             :  * seen here : https://akielaries.github.io/openGPMP/ along with
      19             :  * important files seen in this project.
      20             :  *
      21             :  * You may opt to use, copy, modify, merge, publish, distribute
      22             :  * and/or sell copies of the Software, and permit persons to whom
      23             :  * the Software is furnished to do so, under the terms of the
      24             :  * LICENSE file. As this is an Open Source effort, all implementations
      25             :  * must be of the same methodology.
      26             :  *
      27             :  *
      28             :  *
      29             :  * This software is distributed on an AS IS basis, WITHOUT
      30             :  * WARRANTY OF ANY KIND, either express or implied.
      31             :  *
      32             :  ************************************************************************/
      33             : #include <cmath>
      34             : #include <cstdlib>
      35             : #include <iostream>
      36             : #include <openGPMP/ml/kohonen_net.hpp>
      37             : 
      38           0 : gpmp::ml::KohonenNet::KohonenNet(int in_size, int m_size)
      39           0 :     : input_size(in_size), map_size(m_size) {
      40           0 :     initialize_weights();
      41           0 : }
      42             : 
      43           0 : gpmp::ml::KohonenNet::~KohonenNet() {
      44           0 : }
      45             : 
      46           0 : void gpmp::ml::KohonenNet::initialize_weights() {
      47           0 :     weights.resize(map_size, std::vector<double>(input_size));
      48             : 
      49             :     // Initialize weights randomly or using some other strategy
      50             :     // For simplicity, we initialize them with random values between 0 and 1
      51           0 :     for (int i = 0; i < map_size; ++i) {
      52           0 :         for (int j = 0; j < input_size; ++j) {
      53           0 :             weights[i][j] = static_cast<double>(rand()) / RAND_MAX;
      54             :         }
      55             :     }
      56           0 : }
      57             : 
      58           0 : void gpmp::ml::KohonenNet::train(
      59             :     const std::vector<std::vector<double>> &inputData,
      60             :     int epochs) {
      61           0 :     for (int epoch = 0; epoch < epochs; ++epoch) {
      62           0 :         for (const auto &inputVector : inputData) {
      63           0 :             int bmu = best_matching_unit(inputVector);
      64           0 :             double learningRate = 1.0 - static_cast<double>(epoch) / epochs;
      65           0 :             update_weights(bmu, inputVector, learningRate);
      66             :         }
      67             :     }
      68           0 : }
      69             : 
      70           0 : int gpmp::ml::KohonenNet::classify(const std::vector<double> &inputVector) {
      71           0 :     return best_matching_unit(inputVector);
      72             : }
      73             : 
      74           0 : int gpmp::ml::KohonenNet::best_matching_unit(
      75             :     const std::vector<double> &inputVector) {
      76           0 :     int bmu = 0;
      77           0 :     double minDistance = euclidean_distance(inputVector, weights[0]);
      78             : 
      79           0 :     for (int i = 1; i < map_size; ++i) {
      80           0 :         double distance = euclidean_distance(inputVector, weights[i]);
      81           0 :         if (distance < minDistance) {
      82           0 :             minDistance = distance;
      83           0 :             bmu = i;
      84             :         }
      85             :     }
      86             : 
      87           0 :     return bmu;
      88             : }
      89             : 
      90           0 : void gpmp::ml::KohonenNet::update_weights(
      91             :     int bmu,
      92             :     const std::vector<double> &inputVector,
      93             :     double learningRate) {
      94           0 :     for (int i = 0; i < input_size; ++i) {
      95           0 :         double delta = inputVector[i] - weights[bmu][i];
      96           0 :         weights[bmu][i] += learningRate * delta;
      97             :     }
      98           0 : }
      99             : 
     100             : double
     101           0 : gpmp::ml::KohonenNet::euclidean_distance(const std::vector<double> &vec1,
     102             :                                          const std::vector<double> &vec2) {
     103           0 :     double distance = 0.0;
     104           0 :     for (size_t i = 0; i < vec1.size(); ++i) {
     105           0 :         double diff = vec1[i] - vec2[i];
     106           0 :         distance += diff * diff;
     107             :     }
     108           0 :     return sqrt(distance);
     109             : }