xmmModelSingleClass.hpp

Go to the documentation of this file.

/*
 * xmmModelSingleClass.hpp
 *
 * Abstract class for Single-class Probabilistic Machine learning models based
 * on the EM algorithm
 *
 * Contact:
 * - Jules Francoise <jules.francoise@ircam.fr>
 *
 * This code has been initially authored by Jules Francoise
 * <http://julesfrancoise.com> during his PhD thesis, supervised by Frederic
 * Bevilacqua <href="http://frederic-bevilacqua.net>, in the Sound Music
 * Movement Interaction team <http://ismm.ircam.fr> of the
 * STMS Lab - IRCAM, CNRS, UPMC (2011-2015).
 *
 * Copyright (C) 2015 UPMC, Ircam-Centre Pompidou.
 *
 * This File is part of XMM.
 *
 * XMM is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * XMM is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with XMM.  If not, see <http://www.gnu.org/licenses/>.
 */

#ifndef xmmModelSingleClass_h
#define xmmModelSingleClass_h

#include "../common/xmmCircularbuffer.hpp"
#include "../common/xmmEvents.hpp"
#include "../trainingset/xmmTrainingSet.hpp"
#include "xmmModelSharedParameters.hpp"
#include <memory>
#include <mutex>

namespace xmm {
class TrainingEvent {
  public:
    enum Status {
        Run,

        Error,

        Cancel,

        Done,

        Alldone
    };

    enum class ErrorType {
        ConvergenceError
    };

    TrainingEvent(void* model_, std::string label_,
                  Status status_ = Status::Run)
        : model(model_),
          label(label_),
          status(status_),
          progression(0),
          log_likelihood(0),
          iterations(0) {}

    TrainingEvent(TrainingEvent const& src)
        : model(src.model),
          label(src.label),
          status(src.status),
          progression(src.progression),
          log_likelihood(src.log_likelihood),
          iterations(src.iterations) {}

    TrainingEvent& operator=(TrainingEvent const& src) {
        if (this != &src) {
            model = src.model;
            label = src.label;
            status = src.status;
            progression = src.progression;
            log_likelihood = src.log_likelihood;
            iterations = src.iterations;
        }
        return *this;
    }

    void* model;

    std::string label;

    Status status;

    float progression;

    double log_likelihood;

    double iterations;
};

class SingleClassProbabilisticModel : public Writable {
  public:
    SingleClassProbabilisticModel(std::shared_ptr<SharedParameters> p = NULL);

    SingleClassProbabilisticModel(SingleClassProbabilisticModel const& src);

    explicit SingleClassProbabilisticModel(std::shared_ptr<SharedParameters> p,
                                           Json::Value const& root);

    SingleClassProbabilisticModel& operator=(
        SingleClassProbabilisticModel const& src);

    virtual ~SingleClassProbabilisticModel();


    bool isTraining() const;

    void train(TrainingSet* trainingSet);

    void cancelTraining();



    virtual void reset();

    virtual double filter(std::vector<float> const& observation) = 0;



    virtual Json::Value toJson() const;

    virtual void fromJson(Json::Value const& root) = 0;


    std::string label;

    std::shared_ptr<SharedParameters> shared_parameters;

    EventGenerator<TrainingEvent> training_events;

    TrainingEvent training_status;

  protected:
    virtual void allocate() = 0;

    virtual void emAlgorithmInit(TrainingSet* trainingSet) = 0;

    virtual double emAlgorithmUpdate(TrainingSet* trainingSet) = 0;

    virtual void emAlgorithmTerminate();

    bool emAlgorithmHasConverged(int step, double log_prob,
                                 double old_log_prob) const;

    bool cancelTrainingIfRequested();

    inline void check_training() const {
        if (this->isTraining())
            throw std::runtime_error("The model is training");
    }

    CircularBuffer<double> likelihood_buffer_;

    std::mutex training_mutex_;

    bool is_training_;

    bool cancel_training_;
};
}

#endif