Line data    Source code

       1              : // ***************************************************************
       2              : // SPDX-FileCopyrightText: Copyright 2024 Ricardo Montañana Gómez
       3              : // SPDX-FileType: SOURCE
       4              : // SPDX-License-Identifier: MIT
       5              : // ***************************************************************
       6              : 
       7              : #ifndef BOOSTAODE_H
       8              : #define BOOSTAODE_H
       9              : #include <map>
      10              : #include "bayesnet/classifiers/SPODE.h"
      11              : #include "bayesnet/feature_selection/FeatureSelect.h"
      12              : #include "Ensemble.h"
      13              : namespace bayesnet {
      14              :     const struct {
      15              :         std::string CFS = "CFS";
      16              :         std::string FCBF = "FCBF";
      17              :         std::string IWSS = "IWSS";
      18              :     }SelectFeatures;
      19              :     const struct {
      20              :         std::string ASC = "asc";
      21              :         std::string DESC = "desc";
      22              :         std::string RAND = "rand";
      23              :     }Orders;
      24              :     class BoostAODE : public Ensemble {
      25              :     public:
      26              :         explicit BoostAODE(bool predict_voting = false);
      27           88 :         virtual ~BoostAODE() = default;
      28              :         std::vector<std::string> graph(const std::string& title = "BoostAODE") const override;
      29              :         void setHyperparameters(const nlohmann::json& hyperparameters_) override;
      30              :     protected:
      31              :         void buildModel(const torch::Tensor& weights) override;
      32              :         void trainModel(const torch::Tensor& weights) override;
      33              :     private:
      34              :         std::tuple<torch::Tensor&, double, bool> update_weights_block(int k, torch::Tensor& ytrain, torch::Tensor& weights);
      35              :         std::vector<int> initializeModels();
      36              :         torch::Tensor X_train, y_train, X_test, y_test;
      37              :         // Hyperparameters
      38              :         bool bisection = true; // if true, use bisection stratety to add k models at once to the ensemble
      39              :         int maxTolerance = 3;
      40              :         std::string order_algorithm; // order to process the KBest features asc, desc, rand
      41              :         bool convergence = true; //if true, stop when the model does not improve
      42              :         bool convergence_best = false; // wether to keep the best accuracy to the moment or the last accuracy as prior accuracy
      43              :         bool selectFeatures = false; // if true, use feature selection
      44              :         std::string select_features_algorithm = Orders.DESC; // Selected feature selection algorithm
      45              :         FeatureSelect* featureSelector = nullptr;
      46              :         double threshold = -1;
      47              :         bool block_update = false;
      48              :     };
      49              : }
      50              : #endif