nnfw/src/libradialfunctions.cpp

/********************************************************************************
 *  Neural Network Framework.                                                   *
 *  Copyright (C) 2005-2011 Gianluca Massera <emmegian@yahoo.it>                *
 *                                                                              *
 *  This program 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 2 of the License, or           *
 *  (at your option) any later version.                                         *
 *                                                                              *
 *  This program 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 this program; if not, write to the Free Software                 *
 *  Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA  *
 ********************************************************************************/

#include "libradialfunctions.h"

namespace farsa {

GaussFunction::GaussFunction( double centre, double variance, double maxvalue )
    : OutputFunction() {
    this->centre = centre;
    this->variancev = variance;
    msqrvar = -( variancev*variancev );
    this->max = maxvalue;
}

bool GaussFunction::setVariance( double v ) {
    variancev = v;
    msqrvar = -( variancev*variancev );
    return true;
}

double GaussFunction::variance() {
    return variancev;
}

void GaussFunction::apply( DoubleVector& inputs, DoubleVector& outputs ) {
    // --- out <- max * exp( (centre-inputs)^2 / -(variance^2) )
    square( subtract( outputs, centre, inputs ) );
    exp( outputs /= msqrvar ) *= max;
}

bool GaussFunction::derivate( const DoubleVector& x, const DoubleVector& y, DoubleVector& d ) const {
    // --- d <- ( 2.0*(centre-x) / variance^2 ) * y
    subtract( d, centre, x ) *= 2.0;
    d /= -msqrvar;
    d *= y;
    return true;
}

void GaussFunction::configure(ConfigurationParameters& params, QString prefix)
{
    centre = 0.0;
    QString str = params.getValue(prefix + "centre");
    if (!str.isEmpty()) {
        bool ok;
        centre = str.toDouble(&ok);
        if (!ok) {
            centre = 0.0;
        }
    }

    variancev = 1.0;
    str = params.getValue(prefix + "variance");
    if (!str.isEmpty()) {
        bool ok;
        variancev = str.toDouble(&ok);
        if (!ok) {
            variancev = 1.0;
        }
    }
    // Also recomputing minus squared variance
    msqrvar = -( variancev*variancev );

    max = 1.0;
    str = params.getValue(prefix + "max");
    if (!str.isEmpty()) {
        bool ok;
        max = str.toDouble(&ok);
        if (!ok) {
            max = 1.0;
        }
    }
}

void GaussFunction::save(ConfigurationParameters& params, QString prefix)
{
    params.startObjectParameters(prefix, "GaussFunction", this);
    params.createParameter(prefix, "centre", QString::number(centre));
    params.createParameter(prefix, "variance", QString::number(variancev));
    params.createParameter(prefix, "max", QString::number(max));
}

}