/*
* 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., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
/*
* RuleItem.java
* Copyright (C) 2004 University of Waikato, Hamilton, New Zealand
*
*/
package weka.associations;
import weka.core.Instances;
import weka.core.RevisionHandler;
import weka.core.RevisionUtils;
import java.io.Serializable;
import java.util.Hashtable;
/**
* Class for storing an (class) association rule.
* The premise and the consequence are stored each as separate item sets.
* For every rule their expected predictive accuracy and the time of generation is stored.
* These two measures allow to introduce a sort order for rules.
*
* @author Stefan Mutter
* @version $Revision: 1.5 $
*/
public class RuleItem
implements Comparable, Serializable, RevisionHandler {
/** for serialization */
private static final long serialVersionUID = -3761299128347476534L;
/** The premise of a rule. */
protected ItemSet m_premise;
/** The consequence of a rule. */
protected ItemSet m_consequence;
/** The expected predictive accuracy of a rule. */
protected double m_accuracy;
/** The generation time of a rule. */
protected int m_genTime;
/**
* Constructor for an empty RuleItem
*/
public RuleItem(){
}
/**
* Constructor that generates a RuleItem out of a given one
* @param toCopy RuleItem to copy
*/
public RuleItem(RuleItem toCopy){
m_premise = toCopy.m_premise;
m_consequence = toCopy.m_consequence;
m_accuracy = toCopy.m_accuracy;
m_genTime = toCopy.m_genTime;
}
/**
* Constructor
* @param premise the premise of the future RuleItem
* @param consequence the consequence of the future RuleItem
* @param genTime the time of generation of the future RuleItem
* @param ruleSupport support of the rule
* @param m_midPoints the mid poitns of the intervals
* @param m_priors Hashtable containing the estimated prior probablilities
*/
public RuleItem(ItemSet premise, ItemSet consequence, int genTime,int ruleSupport,double [] m_midPoints, Hashtable m_priors){
m_premise = premise;
m_consequence = consequence;
m_accuracy = RuleGeneration.expectation((double)ruleSupport,m_premise.m_counter,m_midPoints,m_priors);
//overflow, underflow
if(Double.isNaN(m_accuracy) || m_accuracy < 0){
m_accuracy = Double.MIN_VALUE;
}
m_consequence.m_counter = ruleSupport;
m_genTime = genTime;
}
/**
* Constructs a new RuleItem if the support of the given rule is above the support threshold.
* @param premise the premise
* @param consequence the consequence
* @param instances the instances
* @param genTime the time of generation of the current premise and consequence
* @param minRuleCount the support threshold
* @param m_midPoints the mid points of the intervals
* @param m_priors the estimated priori probabilities (in a hashtable)
* @return a RuleItem if its support is above the threshold, null otherwise
*/
public RuleItem generateRuleItem(ItemSet premise, ItemSet consequence, Instances instances,int genTime, int minRuleCount,double[] m_midPoints, Hashtable m_priors){
ItemSet rule = new ItemSet(instances.numInstances());
rule.m_items = new int[(consequence.m_items).length];
System.arraycopy(premise.m_items, 0, rule.m_items, 0, (premise.m_items).length);
for(int k = 0;k < consequence.m_items.length; k++){
if(consequence.m_items[k] != -1)
rule.m_items[k] = consequence.m_items[k];
}
for (int i = 0; i < instances.numInstances(); i++)
rule.upDateCounter(instances.instance(i));
int ruleSupport = rule.support();
if(ruleSupport > minRuleCount){
RuleItem newRule = new RuleItem(premise,consequence,genTime,ruleSupport,m_midPoints,m_priors);
return newRule;
}
return null;
}
//Note: this class has a natural ordering that is inconsistent with equals
/**
* compares two RuleItems and allows an ordering concerning
* expected predictive accuracy and time of generation
* Note: this class has a natural ordering that is inconsistent with equals
* @param o RuleItem to compare
* @return integer indicating the sort oder of the two RuleItems
*/
public int compareTo(Object o) {
if(this.m_accuracy == ((RuleItem)o).m_accuracy){
if((this.m_genTime == ((RuleItem)o).m_genTime))
return 0;
if(this.m_genTime > ((RuleItem)o).m_genTime)
return -1;
if(this.m_genTime < ((RuleItem)o).m_genTime)
return 1;
}
if(this.m_accuracy < ((RuleItem)o).m_accuracy)
return -1;
return 1;
}
/**
* returns whether two RuleItems are equal
* @param o RuleItem to compare
* @return true if the rules are equal, false otherwise
*/
public boolean equals(Object o){
if(o == null)
return false;
if(m_premise.equals(((RuleItem)o).m_premise) && m_consequence.equals(((RuleItem)o).m_consequence))
return true;
return false;
}
/**
* Gets the expected predictive accuracy of a rule
* @return the expected predictive accuracy of a rule stored as a RuleItem
*/
public double accuracy(){
return m_accuracy;
}
/**
* Gets the premise of a rule
* @return the premise of a rule stored as a RuleItem
*/
public ItemSet premise(){
return m_premise;
}
/**
* Gets the consequence of a rule
* @return the consequence of a rule stored as a RuleItem
*/
public ItemSet consequence(){
return m_consequence;
}
/**
* Returns the revision string.
*
* @return the revision
*/
public String getRevision() {
return RevisionUtils.extract("$Revision: 1.5 $");
}
}