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 <algorithm>
      34             : #include <cmath>
      35             : #include <openGPMP/ml/knn.hpp>
      36             : #include <stdexcept>
      37             : #include <unordered_map>
      38             : 
      39           0 : gpmp::ml::KNN::KNN() {
      40           0 : }
      41             : 
      42           0 : gpmp::ml::KNN::~KNN() {
      43           0 : }
      44             : 
      45           0 : void gpmp::ml::KNN::train(
      46             :     const std::vector<std::vector<double>> &_training_data,
      47             :     const std::vector<int> &_labels) {
      48           0 :     if (_training_data.size() != _labels.size() || _training_data.empty()) {
      49           0 :         throw std::invalid_argument("Invalid training data or labels.");
      50             :     }
      51             : 
      52           0 :     this->training_data = _training_data;
      53           0 :     this->labels = _labels;
      54           0 : }
      55             : 
      56           0 : int gpmp::ml::KNN::predict(const std::vector<double> &input_vector, int k) {
      57           0 :     if (training_data.empty() || labels.empty()) {
      58           0 :         throw std::logic_error(
      59           0 :             "Model not trained. Call train() before predict.");
      60             :     }
      61             : 
      62           0 :     if (input_vector.size() != training_data[0].size()) {
      63           0 :         throw std::invalid_argument("Invalid input vector size.");
      64             :     }
      65             : 
      66             :     // if (k <= 0 || k > training_data.size()) {
      67           0 :     if (k <= 0 || static_cast<size_t>(k) > training_data.size()) {
      68           0 :         throw std::invalid_argument("Invalid value of k.");
      69             :     }
      70             : 
      71             :     // Calculate distances and store index-label pairs
      72           0 :     std::vector<std::pair<double, int>> distances;
      73           0 :     for (size_t i = 0; i < training_data.size(); ++i) {
      74             :         double distance =
      75           0 :             calculateEuclideanDistance(input_vector, training_data[i]);
      76           0 :         distances.emplace_back(distance, labels[i]);
      77             :     }
      78             : 
      79             :     // Sort distances in ascending order
      80           0 :     std::sort(
      81             :         distances.begin(),
      82             :         distances.end(),
      83           0 :         [](const std::pair<double, int> &a, const std::pair<double, int> &b) {
      84           0 :             return a.first < b.first;
      85             :         });
      86             : 
      87             :     // Count votes for each label among the k nearest neighbors
      88           0 :     std::unordered_map<int, int> label_counts;
      89           0 :     for (int i = 0; i < k; ++i) {
      90           0 :         label_counts[distances[i].second]++;
      91             :     }
      92             : 
      93             :     // Find the label with the maximum votes
      94           0 :     int max_votes = -1;
      95           0 :     int predicted_label = -1;
      96           0 :     for (const auto &entry : label_counts) {
      97           0 :         if (entry.second > max_votes) {
      98           0 :             max_votes = entry.second;
      99           0 :             predicted_label = entry.first;
     100             :         }
     101             :     }
     102             : 
     103           0 :     return predicted_label;
     104           0 : }
     105             : 
     106             : double
     107           0 : gpmp::ml::KNN::calculateEuclideanDistance(const std::vector<double> &vec1,
     108             :                                           const std::vector<double> &vec2) {
     109           0 :     double distance = 0.0;
     110           0 :     for (size_t i = 0; i < vec1.size(); ++i) {
     111           0 :         double diff = vec1[i] - vec2[i];
     112           0 :         distance += diff * diff;
     113             :     }
     114           0 :     return std::sqrt(distance);
     115             : }