/*
 *  RapidMiner
 *
 *  Copyright (C) 2001-2008 by Rapid-I and the contributors
 *
 *  Complete list of developers available at our web site:
 *
 *       http://rapid-i.com
 *
 *  This program is free software: you can redistribute it and/or modify
 *  it under the terms of the GNU Affero General Public License as published by
 *  the Free Software Foundation, either version 3 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 Affero General Public License for more details.
 *
 *  You should have received a copy of the GNU Affero General Public License
 *  along with this program.  If not, see http://www.gnu.org/licenses/.
 */
package com.rapidminer.operator.learner.igss.utility;

import com.rapidminer.operator.learner.igss.hypothesis.Hypothesis;

/** Abstract superclass for all instance-averaging functions.
 * 
 *  @author Dirk Dach 
 *  @version $Id: InstanceAveraging.java,v 1.3 2008/05/09 19:23:24 ingomierswa Exp $
 */
public abstract class InstanceAveraging extends AbstractUtility{
	
	/** Constructor */
	public InstanceAveraging(double[] priors, int large) {
		super(priors,large);
	}
	
	/** Calculates the the confidence intervall for a specific hypothesis. 
	 *  Uses Chernoff bounds if the number of random experiments is too small and normal approximation otherwise.
	 *  This method is adapted for instance averaging utility types. Every example is considered a random experiment, because
	 *  f_inst is evaluated for every example!!! This is the reason why total weight is used instead of covered weight
	 *  Should be overwritten by subclasses if they make a different random experiment.*/
	public double confidenceIntervall (double totalWeight, double totalPositiveWeight, Hypothesis hypo, double delta) {
		if (totalWeight<large) {
			return confSmallM(totalWeight,delta);
		}
		else {
			return conf(totalWeight,totalPositiveWeight,hypo,delta);
		}
	}
	
	/** Calculate confidence intervall without a specific rule for instance averaging functions.*/
	public double conf (double totalWeight, double delta) {
		return inverseNormal(1-delta/2)/(2*Math.sqrt(totalWeight));
	}
	
	/** Calculate confidence intervall for a specific rule for instance averaging functions.*/
	public double conf (double totalWeight, double totalPositiveWeight, Hypothesis hypo, double delta) {
		return inverseNormal(1-delta/2)*variance(totalWeight,totalPositiveWeight,hypo);
	}
	
	/** Calculates the empirical variance. */
	public abstract double variance(double totalWeight, double totalPositiveWeight, Hypothesis hypo);
	
	/** Calculate confidence intervall without a specific rule for instance averaging functions and small m. */
	public double confSmallM (double totalWeight, double delta) {
		return Math.sqrt(Math.log(2.0d/delta)/(2*totalWeight));
	}
}