/*
* 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.
*/
/*
* Tertius.java
* Copyright (C) 2003 Peter A. Flach, Nicolas Lachiche
*
* Thanks to Amelie Deltour for porting the original C code to Java
* and integrating it into Weka.
*/
package weka.associations;
import weka.associations.tertius.*;
import weka.core.*;
import weka.core.Capabilities.Capability;
import weka.core.TechnicalInformation.Field;
import weka.core.TechnicalInformation.Type;
import java.awt.*;
import java.awt.event.ActionEvent;
import java.awt.event.ActionListener;
import java.io.BufferedReader;
import java.io.File;
import java.io.FileReader;
import java.io.Reader;
import java.util.ArrayList;
import java.util.Date;
import java.util.Enumeration;
import java.util.Vector;
/**
<!-- globalinfo-start -->
* Finds rules according to confirmation measure (Tertius-type algorithm).<br/>
* <br/>
* For more information see:<br/>
* <br/>
* P. A. Flach, N. Lachiche (1999). Confirmation-Guided Discovery of first-order rules with Tertius. Machine Learning. 42:61-95.
* <p/>
<!-- globalinfo-end -->
*
<!-- technical-bibtex-start -->
* BibTeX:
* <pre>
* @article{Flach1999,
* author = {P. A. Flach and N. Lachiche},
* journal = {Machine Learning},
* pages = {61-95},
* title = {Confirmation-Guided Discovery of first-order rules with Tertius},
* volume = {42},
* year = {1999}
* }
* </pre>
* <p/>
<!-- technical-bibtex-end -->
*
<!-- options-start -->
* Valid options are: <p/>
*
* <pre> -K <number of values in result>
* Set maximum number of confirmation values in the result. (default: 10)</pre>
*
* <pre> -F <frequency threshold>
* Set frequency threshold for pruning. (default: 0)</pre>
*
* <pre> -C <confirmation threshold>
* Set confirmation threshold. (default: 0)</pre>
*
* <pre> -N <noise threshold>
* Set noise threshold : maximum frequency of counter-examples.
* 0 gives only satisfied rules. (default: 1)</pre>
*
* <pre> -R
* Allow attributes to be repeated in a same rule.</pre>
*
* <pre> -L <number of literals>
* Set maximum number of literals in a rule. (default: 4)</pre>
*
* <pre> -G <0=no negation | 1=body | 2=head | 3=body and head>
* Set the negations in the rule. (default: 0)</pre>
*
* <pre> -S
* Consider only classification rules.</pre>
*
* <pre> -c <class index>
* Set index of class attribute. (default: last).</pre>
*
* <pre> -H
* Consider only horn clauses.</pre>
*
* <pre> -E
* Keep equivalent rules.</pre>
*
* <pre> -M
* Keep same clauses.</pre>
*
* <pre> -T
* Keep subsumed rules.</pre>
*
* <pre> -I <0=always match | 1=never match | 2=significant>
* Set the way to handle missing values. (default: 0)</pre>
*
* <pre> -O
* Use ROC analysis. </pre>
*
* <pre> -p <name of file>
* Set the file containing the parts of the individual for individual-based learning.</pre>
*
* <pre> -P <0=no output | 1=on stdout | 2=in separate window>
* Set output of current values. (default: 0)</pre>
*
<!-- options-end -->
*
* @author <a href="mailto:adeltour@netcourrier.com">Amelie Deltour</a>
* @version $Revision: 5444 $
*/
public class Tertius
extends AbstractAssociator
implements OptionHandler, Runnable, TechnicalInformationHandler {
/** for serialization */
static final long serialVersionUID = 5556726848380738179L;
/** The results. */
private SimpleLinkedList m_results;
/** Number of hypotheses considered. */
private int m_hypotheses;
/** Number of hypotheses explored. */
private int m_explored;
/** Time needed for the search. */
private Date m_time;
/** Field to output the current values. */
private TextField m_valuesText;
/** Instances used for the search. */
private Instances m_instances;
/** Predicates used in the rules. */
private ArrayList m_predicates;
/** Status of the search. */
private int m_status;
/** Status of the search: normal */
private static final int NORMAL = 0;
/** Status of the search: memory problem */
private static final int MEMORY = 1;
/** Status of the search: user interruption */
private static final int STOP = 2;
/* Pruning options. */
/** Number of best confirmation values to search. */
private int m_best;
/** Frequency threshold for the body and the negation of the head. */
private double m_frequencyThreshold;
/** Confirmation threshold for the rules. */
private double m_confirmationThreshold;
/** Maximal number of counter-instances. */
private double m_noiseThreshold;
/* Search space & language bias options. */
/** Repeat attributes ? */
private boolean m_repeat;
/** Number of literals in a rule. */
private int m_numLiterals;
/** Type of negation: none */
private static final int NONE = 0;
/** Type of negation: body */
private static final int BODY = 1;
/** Type of negation: head */
private static final int HEAD = 2;
/** Type of negation: all */
private static final int ALL = 3;
/** Types of negation. */
private static final Tag [] TAGS_NEGATION = {
new Tag(NONE, "None"),
new Tag(BODY, "Body"),
new Tag(HEAD, "Head"),
new Tag(ALL, "Both")
};
/** Type of negation used in the rules. */
private int m_negation;
/** Classification bias. */
private boolean m_classification;
/** Index of class attribute. */
private int m_classIndex;
/** Horn clauses bias. */
private boolean m_horn;
/* Subsumption tests options. */
/** Perform test on equivalent rules ? */
private boolean m_equivalent;
/** Perform test on same clauses ? */
private boolean m_sameClause;
/** Perform subsumption test ? */
private boolean m_subsumption;
/** Way of handling missing values: min counterinstances */
public static final int EXPLICIT = 0;
/** Way of handling missing values: max counterinstances */
public static final int IMPLICIT = 1;
/** Way of handling missing values: missing as a particular value */
public static final int SIGNIFICANT = 2;
/** Ways of handling missing values. */
private static final Tag [] TAGS_MISSING = {
new Tag(EXPLICIT, "Matches all"),
new Tag(IMPLICIT, "Matches none"),
new Tag(SIGNIFICANT, "Significant")
};
/** Way of handling missing values in the search. */
private int m_missing;
/** Perform ROC analysis ? */
private boolean m_roc;
/** Name of the file containing the parts for individual-based learning. */
private String m_partsString;
/** Part instances for individual-based learning. */
private Instances m_parts;
/** Type of values output: No */
private static final int NO = 0;
/** Type of values output: stdout */
private static final int OUT = 1;
/** Type of values output: Window */
private static final int WINDOW = 2;
/** Types of values output. */
private static final Tag [] TAGS_VALUES = {
new Tag(NO, "No"),
new Tag(OUT, "stdout"),
new Tag(WINDOW, "Window")
};
/** Type of values output. */
private int m_printValues;
/**
* Constructor that sets the options to the default values.
*/
public Tertius() {
resetOptions();
}
/**
* Returns a string describing this associator.
*
* @return A description of the evaluator suitable for
* displaying in the explorer/experimenter gui.
*/
public String globalInfo() {
return
"Finds rules according to confirmation measure (Tertius-type "
+ "algorithm).\n\n"
+ "For more information see:\n\n"
+ getTechnicalInformation().toString();
}
/**
* Returns an instance of a TechnicalInformation object, containing
* detailed information about the technical background of this class,
* e.g., paper reference or book this class is based on.
*
* @return the technical information about this class
*/
public TechnicalInformation getTechnicalInformation() {
TechnicalInformation result;
result = new TechnicalInformation(Type.ARTICLE);
result.setValue(Field.AUTHOR, "P. A. Flach and N. Lachiche");
result.setValue(Field.YEAR, "1999");
result.setValue(Field.TITLE, "Confirmation-Guided Discovery of first-order rules with Tertius");
result.setValue(Field.JOURNAL, "Machine Learning");
result.setValue(Field.VOLUME, "42");
result.setValue(Field.PAGES, "61-95");
return result;
}
/**
* Resets the options to the default values.
*/
public void resetOptions() {
/* Pruning options. */
m_best = 10;
m_frequencyThreshold = 0;
m_confirmationThreshold = 0;
m_noiseThreshold = 1;
/* Search space & language bias options. */
m_repeat = false;
m_numLiterals = 4;
m_negation = NONE;
m_classification = false;
m_classIndex = 0;
m_horn = false;
/* Subsumption tests options. */
m_equivalent = true;
m_sameClause = true;
m_subsumption = true;
/* Missing values. */
m_missing = EXPLICIT;
/* ROC analysis. */
m_roc = false;
/* Individual-based learning. */
m_partsString = "";
m_parts = null;
/* Values output. */
m_printValues = NO;
}
/**
* Returns an enumeration describing the available options.
*
* @return An enumeration of all the available options.
*/
public Enumeration listOptions() {
Vector newVector = new Vector(17);
/* Pruning options. */
newVector.addElement(new Option("\tSet maximum number of confirmation "
+ "values in the result. (default: 10)",
"K", 1, "-K <number of values in result>"));
newVector.addElement(new Option("\tSet frequency threshold for pruning. "
+ "(default: 0)",
"F", 1, "-F <frequency threshold>"));
newVector.addElement(new Option("\tSet confirmation threshold. "
+ "(default: 0)",
"C", 1, "-C <confirmation threshold>"));
newVector.addElement(new Option("\tSet noise threshold : maximum frequency "
+ "of counter-examples.\n\t0 gives only "
+ "satisfied rules. (default: 1)",
"N", 1, "-N <noise threshold>"));
/* Search space & language bias options. */
newVector.addElement(new Option("\tAllow attributes to be repeated in a "
+ "same rule.",
"R", 0, "-R"));
newVector.addElement(new Option("\tSet maximum number of literals in a "
+ "rule. (default: 4)",
"L", 1, "-L <number of literals>"));
newVector.addElement(new Option("\tSet the negations in the rule. "
+ "(default: 0)",
"G", 1, "-G <0=no negation | "
+ "1=body | "
+ "2=head | "
+ "3=body and head>"));
newVector.addElement(new Option("\tConsider only classification rules.",
"S", 0, "-S"));
newVector.addElement(new Option("\tSet index of class attribute. "
+ "(default: last).",
"c", 1, "-c <class index>"));
newVector.addElement(new Option("\tConsider only horn clauses.",
"H", 0, "-H"));
/* Subsumption tests options. */
newVector.addElement(new Option("\tKeep equivalent rules.",
"E", 0, "-E"));
newVector.addElement(new Option("\tKeep same clauses.",
"M", 0, "-M"));
newVector.addElement(new Option("\tKeep subsumed rules.",
"T", 0, "-T"));
/* Missing values options. */
newVector.addElement(new Option("\tSet the way to handle missing values. "
+ "(default: 0)",
"I", 1, "-I <0=always match | "
+ "1=never match | "
+ "2=significant>"));
/* ROC analysis. */
newVector.addElement(new Option("\tUse ROC analysis. ",
"O", 0, "-O"));
/* Individual-based learning. */
newVector.addElement(new Option("\tSet the file containing the parts of "
+ "the individual for individual-based "
+ "learning.",
"p", 1, "-p <name of file>"));
/* Values output. */
newVector.addElement(new Option("\tSet output of current values. "
+ "(default: 0)",
"P", 1, "-P <0=no output | "
+ "1=on stdout | "
+ "2=in separate window>"));
return newVector.elements();
}
/**
* Parses a given list of options. <p/>
*
<!-- options-start -->
* Valid options are: <p/>
*
* <pre> -K <number of values in result>
* Set maximum number of confirmation values in the result. (default: 10)</pre>
*
* <pre> -F <frequency threshold>
* Set frequency threshold for pruning. (default: 0)</pre>
*
* <pre> -C <confirmation threshold>
* Set confirmation threshold. (default: 0)</pre>
*
* <pre> -N <noise threshold>
* Set noise threshold : maximum frequency of counter-examples.
* 0 gives only satisfied rules. (default: 1)</pre>
*
* <pre> -R
* Allow attributes to be repeated in a same rule.</pre>
*
* <pre> -L <number of literals>
* Set maximum number of literals in a rule. (default: 4)</pre>
*
* <pre> -G <0=no negation | 1=body | 2=head | 3=body and head>
* Set the negations in the rule. (default: 0)</pre>
*
* <pre> -S
* Consider only classification rules.</pre>
*
* <pre> -c <class index>
* Set index of class attribute. (default: last).</pre>
*
* <pre> -H
* Consider only horn clauses.</pre>
*
* <pre> -E
* Keep equivalent rules.</pre>
*
* <pre> -M
* Keep same clauses.</pre>
*
* <pre> -T
* Keep subsumed rules.</pre>
*
* <pre> -I <0=always match | 1=never match | 2=significant>
* Set the way to handle missing values. (default: 0)</pre>
*
* <pre> -O
* Use ROC analysis. </pre>
*
* <pre> -p <name of file>
* Set the file containing the parts of the individual for individual-based learning.</pre>
*
* <pre> -P <0=no output | 1=on stdout | 2=in separate window>
* Set output of current values. (default: 0)</pre>
*
<!-- options-end -->
*
* @param options The list of options as an array of strings.
* @throws Exception if an option is not supported.
*/
public void setOptions(String [] options) throws Exception {
resetOptions();
/* Pruning options. */
String bestString = Utils.getOption('K', options);
if (bestString.length() != 0) {
try {
m_best = Integer.parseInt(bestString);
} catch (Exception e) {
throw new Exception("Invalid value for -K option: "
+ e.getMessage() + ".");
}
if (m_best < 1) {
throw new Exception("Number of confirmation values has to be "
+ "greater than one!");
}
}
String frequencyThresholdString = Utils.getOption('F', options);
if (frequencyThresholdString.length() != 0) {
try {
m_frequencyThreshold
= (new Double(frequencyThresholdString)).doubleValue();
} catch (Exception e) {
throw new Exception("Invalid value for -F option: "
+ e.getMessage() + ".");
}
if (m_frequencyThreshold < 0 || m_frequencyThreshold > 1) {
throw new Exception("Frequency threshold has to be between "
+ "zero and one!");
}
}
String confirmationThresholdString = Utils.getOption('C', options);
if (confirmationThresholdString.length() != 0) {
try {
m_confirmationThreshold
= (new Double(confirmationThresholdString)).doubleValue();
} catch (Exception e) {
throw new Exception("Invalid value for -C option: "
+ e.getMessage() + ".");
}
if (m_confirmationThreshold < 0 || m_confirmationThreshold > 1) {
throw new Exception("Confirmation threshold has to be between "
+ "zero and one!");
}
if (bestString.length() != 0) {
throw new Exception("Specifying both a number of confirmation "
+ "values and a confirmation threshold "
+ "doesn't make sense!");
}
if (m_confirmationThreshold != 0) {
m_best = 0;
}
}
String noiseThresholdString = Utils.getOption('N', options);
if (noiseThresholdString.length() != 0) {
try {
m_noiseThreshold = (new Double(noiseThresholdString)).doubleValue();
} catch (Exception e) {
throw new Exception("Invalid value for -N option: "
+ e.getMessage() + ".");
}
if (m_noiseThreshold < 0 || m_noiseThreshold > 1) {
throw new Exception("Noise threshold has to be between "
+ "zero and one!");
}
}
/* Search space and language bias options. */
m_repeat = Utils.getFlag('R', options);
String numLiteralsString = Utils.getOption('L', options);
if (numLiteralsString.length() != 0) {
try {
m_numLiterals = Integer.parseInt(numLiteralsString);
} catch (Exception e) {
throw new Exception("Invalid value for -L option: "
+ e.getMessage() + ".");
}
if (m_numLiterals < 1) {
throw new Exception("Number of literals has to be "
+ "greater than one!");
}
}
String negationString = Utils.getOption('G', options);
if (negationString.length() != 0) {
SelectedTag selected;
int tag;
try {
tag = Integer.parseInt(negationString);
} catch (Exception e) {
throw new Exception("Invalid value for -G option: "
+ e.getMessage() + ".");
}
try {
selected = new SelectedTag(tag, TAGS_NEGATION);
} catch (Exception e) {
throw new Exception("Value for -G option has to be "
+ "between zero and three!");
}
setNegation(selected);
}
m_classification = Utils.getFlag('S', options);
String classIndexString = Utils.getOption('c', options);
if (classIndexString.length() != 0) {
try {
m_classIndex = Integer.parseInt(classIndexString);
} catch (Exception e) {
throw new Exception("Invalid value for -c option: "
+ e.getMessage() + ".");
}
}
m_horn = Utils.getFlag('H', options);
if (m_horn && (m_negation != NONE)) {
throw new Exception("Considering horn clauses doesn't make sense "
+ "if negation allowed!");
}
/* Subsumption tests options. */
m_equivalent = !(Utils.getFlag('E', options));
m_sameClause = !(Utils.getFlag('M', options));
m_subsumption = !(Utils.getFlag('T', options));
/* Missing values options. */
String missingString = Utils.getOption('I', options);
if (missingString.length() != 0) {
SelectedTag selected;
int tag;
try {
tag = Integer.parseInt(missingString);
} catch (Exception e) {
throw new Exception("Invalid value for -I option: "
+ e.getMessage() + ".");
}
try {
selected = new SelectedTag(tag, TAGS_MISSING);
} catch (Exception e) {
throw new Exception("Value for -I option has to be "
+ "between zero and two!");
}
setMissingValues(selected);
}
/* ROC analysis. */
m_roc = Utils.getFlag('O', options);
/* Individual-based learning. */
m_partsString = Utils.getOption('p', options);
if (m_partsString.length() != 0) {
Reader reader;
try {
reader = new BufferedReader(new FileReader(m_partsString));
} catch (Exception e) {
throw new Exception("Can't open file " + e.getMessage() + ".");
}
m_parts = new Instances(reader);
}
/* Values output. */
String printValuesString = Utils.getOption('P', options);
if (printValuesString.length() != 0) {
SelectedTag selected;
int tag;
try {
tag = Integer.parseInt(printValuesString);
} catch (Exception e) {
throw new Exception("Invalid value for -P option: "
+ e.getMessage() + ".");
}
try {
selected = new SelectedTag(tag, TAGS_VALUES);
} catch (Exception e) {
throw new Exception("Value for -P option has to be "
+ "between zero and two!");
}
setValuesOutput(selected);
}
}
/**
* Gets the current settings of the Tertius object.
*
* @return An array of strings suitable for passing to setOptions.
*/
public String [] getOptions() {
Vector result;
result = new Vector();
/* Pruning options. */
if (m_best > 0) {
result.add("-K");
result.add("" + m_best);
}
result.add("-F");
result.add("" + m_frequencyThreshold);
if (m_confirmationThreshold > 0) {
result.add("-C");
result.add("" + m_confirmationThreshold);
}
result.add("-N");
result.add("" + m_noiseThreshold);
/* Search space and language bias options. */
if (m_repeat)
result.add("-R");
result.add("-L");
result.add("" + m_numLiterals);
result.add("-G");
result.add("" + m_negation);
if (m_classification)
result.add("-S");
result.add("-c");
result.add("" + m_classIndex);
if (m_horn)
result.add("-H");
/* Subsumption tests options. */
if (!m_equivalent)
result.add("-E");
if (!m_sameClause)
result.add("-M");
if (!m_subsumption)
result.add("-T");
/* Missing values options. */
result.add("-I");
result.add("" + m_missing);
/* ROC analysis. */
if (m_roc)
result.add("-O");
/* Individual-based learning. */
if (m_partsString.length() > 0) {
result.add("-p");
result.add("" + m_partsString);
}
/* Values output. */
result.add("-P");
result.add("" + m_printValues);
return (String[]) result.toArray(new String[result.size()]);
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String confirmationValuesTipText() {
return "Number of best confirmation values to find.";
}
/**
* Get the value of confirmationValues.
*
* @return Value of confirmationValues.
*/
public int getConfirmationValues() {
return m_best;
}
/**
* Set the value of confirmationValues.
*
* @param v Value to assign to confirmationValues.
*/
public void setConfirmationValues(int v) {
m_best = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String frequencyThresholdTipText() {
return "Minimum proportion of instances satisfying head and body of rules";
}
/**
* Get the value of frequencyThreshold.
*
* @return Value of frequencyThreshold.
*/
public double getFrequencyThreshold() {
return m_frequencyThreshold;
}
/**
* Set the value of frequencyThreshold.
*
* @param v Value to assign to frequencyThreshold.
*/
public void setFrequencyThreshold(double v) {
m_frequencyThreshold = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String confirmationThresholdTipText() {
return "Minimum confirmation of the rules.";
}
/**
* Get the value of confirmationThreshold.
*
* @return Value of confirmationThreshold.
*/
public double getConfirmationThreshold() {
return m_confirmationThreshold;
}
/**
* Set the value of confirmationThreshold.
*
* @param v Value to assign to confirmationThreshold.
*/
public void setConfirmationThreshold(double v) {
m_confirmationThreshold = v;
if (v != 0) {
m_best = 0;
}
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String noiseThresholdTipText() {
return "Maximum proportion of counter-instances of rules. "
+ "If set to 0, only satisfied rules will be given.";
}
/**
* Get the value of noiseThreshold.
*
* @return Value of noiseThreshold.
*/
public double getNoiseThreshold() {
return m_noiseThreshold;
}
/**
* Set the value of noiseThreshold.
*
* @param v Value to assign to noiseThreshold.
*/
public void setNoiseThreshold(double v) {
m_noiseThreshold = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String repeatLiteralsTipText() {
return "Repeated attributes allowed.";
}
/**
* Get the value of repeatLiterals.
*
* @return Value of repeatLiterals.
*/
public boolean getRepeatLiterals() {
return m_repeat;
}
/**
* Set the value of repeatLiterals.
*
* @param v Value to assign to repeatLiterals.
*/
public void setRepeatLiterals(boolean v) {
m_repeat = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String numberLiteralsTipText() {
return "Maximum number of literals in a rule.";
}
/**
* Get the value of numberLiterals.
*
* @return Value of numberLiterals.
*/
public int getNumberLiterals() {
return m_numLiterals;
}
/**
* Set the value of numberLiterals.
*
* @param v Value to assign to numberLiterals.
*/
public void setNumberLiterals(int v) {
m_numLiterals = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String negationTipText() {
return "Set the type of negation allowed in the rule. "
+ "Negation can be allowed in the body, in the head, in both "
+ "or in none.";
}
/**
* Get the value of negation.
*
* @return Value of negation.
*/
public SelectedTag getNegation() {
return new SelectedTag(m_negation, TAGS_NEGATION);
}
/**
* Set the value of negation.
*
* @param v Value to assign to negation.
*/
public void setNegation(SelectedTag v) {
if (v.getTags() == TAGS_NEGATION) {
m_negation = v.getSelectedTag().getID();
}
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String classificationTipText() {
return "Find only rules with the class in the head.";
}
/**
* Get the value of classification.
*
* @return Value of classification.
*/
public boolean getClassification() {
return m_classification;
}
/**
* Set the value of classification.
*
* @param v Value to assign to classification.
*/
public void setClassification(boolean v) {
m_classification = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String classIndexTipText() {
return "Index of the class attribute. If set to 0, the class will be the last attribute.";
}
/**
* Get the value of classIndex.
*
* @return Value of classIndex.
*/
public int getClassIndex() {
return m_classIndex;
}
/**
* Set the value of classIndex.
*
* @param v Value to assign to classIndex.
*/
public void setClassIndex(int v) {
m_classIndex = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String hornClausesTipText() {
return "Find rules with a single conclusion literal only.";
}
/**
* Get the value of hornClauses.
*
* @return Value of hornClauses.
*/
public boolean getHornClauses() {
return m_horn;
}
/**
* Set the value of hornClauses.
*
* @param v Value to assign to hornClauses.
*/
public void setHornClauses(boolean v) {
m_horn = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String equivalentTipText() {
return "Keep equivalent rules. "
+ "A rule r2 is equivalent to a rule r1 if the body of r2 is the "
+ "negation of the head of r1, and the head of r2 is the "
+ "negation of the body of r1.";
}
/**
* Get the value of equivalent.
*
* @return Value of equivalent.
*/
public boolean disabled_getEquivalent() {
return !m_equivalent;
}
/**
* Set the value of equivalent.
*
* @param v Value to assign to equivalent.
*/
public void disabled_setEquivalent(boolean v) {
m_equivalent = !v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String sameClauseTipText() {
return "Keep rules corresponding to the same clauses. "
+ "If set to false, only the rule with the best confirmation "
+ "value and rules with a lower number of counter-instances "
+ "will be kept.";
}
/**
* Get the value of sameClause.
*
* @return Value of sameClause.
*/
public boolean disabled_getSameClause() {
return !m_sameClause;
}
/**
* Set the value of sameClause.
*
* @param v Value to assign to sameClause.
*/
public void disabled_setSameClause(boolean v) {
m_sameClause = !v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String subsumptionTipText() {
return "Keep subsumed rules. "
+ "If set to false, subsumed rules will only be kept if they "
+ "have a better confirmation or a lower number of counter-instances.";
}
/**
* Get the value of subsumption.
*
* @return Value of subsumption.
*/
public boolean disabled_getSubsumption() {
return !m_subsumption;
}
/**
* Set the value of subsumption.
*
* @param v Value to assign to subsumption.
*/
public void disabled_setSubsumption(boolean v) {
m_subsumption = !v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String missingValuesTipText() {
return "Set the way to handle missing values. "
+ "Missing values can be set to match any value, or never match values "
+ "or to be significant and possibly appear in rules.";
}
/**
* Get the value of missingValues.
*
* @return Value of missingValues.
*/
public SelectedTag getMissingValues() {
return new SelectedTag(m_missing, TAGS_MISSING);
}
/**
* Set the value of missingValues.
*
* @param v Value to assign to missingValues.
*/
public void setMissingValues(SelectedTag v) {
if (v.getTags() == TAGS_MISSING) {
m_missing = v.getSelectedTag().getID();
}
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String rocAnalysisTipText() {
return "Return TP-rate and FP-rate for each rule found.";
}
/**
* Get the value of rocAnalysis.
*
* @return Value of rocAnalysis.
*/
public boolean getRocAnalysis() {
return m_roc;
}
/**
* Set the value of rocAnalysis.
*
* @param v Value to assign to rocAnalysis.
*/
public void setRocAnalysis(boolean v) {
m_roc = v;
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String partFileTipText() {
return "Set file containing the parts of the individual "
+ "for individual-based learning.";
}
/**
* Get the value of partFile.
*
* @return Value of partFile.
*/
public File disabled_getPartFile() {
return new File(m_partsString);
}
/**
* Set the value of partFile.
*
* @param v Value to assign to partFile.
* @throws Exception if file cannot be opened
*/
public void disabled_setPartFile(File v) throws Exception {
m_partsString = v.getAbsolutePath();
if (m_partsString.length() != 0) {
Reader reader;
try {
reader = new BufferedReader(new FileReader(m_partsString));
} catch (Exception e) {
throw new Exception("Can't open file " + e.getMessage() + ".");
}
m_parts = new Instances(reader);
}
}
/**
* Returns the tip text for this property.
*
* @return Tip text for this property suitable for
* displaying in the explorer/experimenter GUI.
*/
public String valuesOutputTipText() {
return "Give visual feedback during the search. "
+ "The current best and worst values can be output either to stdout or to a separate window.";
}
/**
* Get the value of valuesOutput.
*
* @return Value of valuesOutput.
*/
public SelectedTag getValuesOutput() {
return new SelectedTag(m_printValues, TAGS_VALUES);
}
/**
* Set the value of valuesOutput.
*
* @param v Value to assign to valuesOutput.
*/
public void setValuesOutput(SelectedTag v) {
if (v.getTags() == TAGS_VALUES) {
m_printValues = v.getSelectedTag().getID();
}
}
/**
* Build the predicate corresponding to an attribute.
*
* @param instances The instances this predicates comes from.
* @param attr The attribute this predicate corresponds to.
* @param isClass Saying if the attribute is the class attribute.
* @return The corresponding Predicate.
* @throws Exception if the predicate could not be build
* (the attribute is numeric).
*/
private Predicate buildPredicate(Instances instances,
Attribute attr, boolean isClass)
throws Exception {
Predicate predicate; /* The result. */
Literal lit;
Literal negation;
boolean missingValues; /* Missing values for this attribute ? */
boolean individual = (m_parts != null); /* Individual-based learning ? */
int type = (instances == m_parts)
? IndividualLiteral.PART_PROPERTY
: IndividualLiteral.INDIVIDUAL_PROPERTY; /* Type of property. */
if (attr.isNumeric()) {
throw new Exception("Can't handle numeric attributes!");
}
missingValues = instances.attributeStats(attr.index()).missingCount > 0;
/* Build predicate. */
if (individual) {
predicate = new Predicate(instances.relationName() + "." + attr.name(),
attr.index(), isClass);
} else {
predicate = new Predicate(attr.name(), attr.index(), isClass);
}
if (attr.numValues() == 2
&& (!missingValues || m_missing == EXPLICIT)) {
/* Case of two values.
* If there are missing values, this case is treated like other cases.
*/
if (individual) {
lit = new IndividualLiteral(predicate, attr.value(0), 0,
Literal.POS, m_missing, type);
negation = new IndividualLiteral(predicate, attr.value(1), 1,
Literal.POS, m_missing, type);
} else {
lit = new AttributeValueLiteral(predicate, attr.value(0), 0,
Literal.POS, m_missing);
negation = new AttributeValueLiteral(predicate, attr.value(1), 1,
Literal.POS, m_missing);
}
lit.setNegation(negation);
negation.setNegation(lit);
predicate.addLiteral(lit);
} else {
/* Case of several values. */
for (int i = 0; i < attr.numValues(); i++) {
if (individual) {
lit = new IndividualLiteral(predicate, attr.value(i), i,
Literal.POS, m_missing, type);
} else {
lit = new AttributeValueLiteral(predicate, attr.value(i), i,
Literal.POS, m_missing);
}
if (m_negation != NONE) {
if (individual) {
negation = new IndividualLiteral(predicate, attr.value(i), i,
Literal.NEG, m_missing, type);
} else {
negation = new AttributeValueLiteral(predicate, attr.value(i), i,
Literal.NEG, m_missing);
}
lit.setNegation(negation);
negation.setNegation(lit);
}
predicate.addLiteral(lit);
}
/* One more value if missing is significant. */
if (missingValues && m_missing == SIGNIFICANT) {
if (individual) {
lit = new IndividualLiteral(predicate, "?", -1,
Literal.POS, m_missing, type);
} else {
lit = new AttributeValueLiteral(predicate, "?", -1,
Literal.POS, m_missing);
}
if (m_negation != NONE) {
if (individual) {
negation = new IndividualLiteral(predicate, "?", -1,
Literal.NEG, m_missing, type);
} else {
negation = new AttributeValueLiteral(predicate, "?", -1,
Literal.NEG, m_missing);
}
lit.setNegation(negation);
negation.setNegation(lit);
}
predicate.addLiteral(lit);
}
}
return predicate;
}
/**
* Build the predicates to use in the rules.
*
* @return the predicates
* @throws Exception If the predicates could not be built
* (numeric attribute).
*/
private ArrayList buildPredicates() throws Exception {
ArrayList predicates = new ArrayList(); /* The result. */
Predicate predicate;
Attribute attr;
Enumeration attributes = m_instances.enumerateAttributes();
boolean individual = (m_parts != null); /* Individual-based learning ? */
/* Attributes. */
while (attributes.hasMoreElements()) {
attr = (Attribute) attributes.nextElement();
/* Identifiers do not appear in rules in individual-based learning. */
if (!(individual && attr.name().equals("id"))) {
predicate = buildPredicate(m_instances, attr, false);
predicates.add(predicate);
}
}
/* Class attribute. */
attr = m_instances.classAttribute();
/* Identifiers do not appear in rules. */
if (!(individual && attr.name().equals("id"))) {
predicate = buildPredicate(m_instances, attr, true);
predicates.add(predicate);
}
/* Attributes of the parts in individual-based learning. */
if (individual) {
attributes = m_parts.enumerateAttributes();
while (attributes.hasMoreElements()) {
attr = (Attribute) attributes.nextElement();
/* Identifiers do not appear in rules. */
if (!attr.name().equals("id")) {
predicate = buildPredicate(m_parts, attr, false);
predicates.add(predicate);
}
}
}
return predicates;
}
/**
* Count the number of distinct confirmation values in the results.
*
* @return Number of confirmation values in the results.
*/
private int numValuesInResult() {
int result = 0;
SimpleLinkedList.LinkedListIterator iter = m_results.iterator();
Rule current;
Rule next;
if (!iter.hasNext()) {
return result;
} else {
current = (Rule) iter.next();
while (iter.hasNext()) {
next = (Rule) iter.next();
if (current.getConfirmation() > next.getConfirmation()) {
result++;
}
current = next;
}
return result + 1;
}
}
/**
* Test if it is worth refining a rule.
*
* @param rule The rule to consider.
* @return True if the rule can be refined, false otherwise.
*/
private boolean canRefine(Rule rule) {
if (rule.isEmpty()) {
return true;
}
if (m_best != 0) {
if (numValuesInResult() < m_best) {
return true;
}
Rule worstResult = (Rule) m_results.getLast();
if (rule.getOptimistic() >= worstResult.getConfirmation()) {
return true;
}
return false;
} else {
return true;
}
}
/**
* Test if it is worth calculating the optimistic estimate of a rule.
*
* @param rule The rule to consider.
* @return True if the optimistic estimate can be calculated, false otherwise.
*/
private boolean canCalculateOptimistic(Rule rule) {
if (rule.hasTrueBody() || rule.hasFalseHead()) {
return false;
}
if (!rule.overFrequencyThreshold(m_frequencyThreshold)) {
return false;
}
return true;
}
/**
* Test if a rule can be explored (if it is interesting for the results
* or for refining).
*
* @param rule The rule to consider.
* @return True if the rule can be explored, false otherwise.
*/
private boolean canExplore(Rule rule) {
if (rule.getOptimistic() < m_confirmationThreshold) {
return false;
}
if (m_best != 0) {
if (numValuesInResult() < m_best) {
return true;
}
Rule worstResult = (Rule) m_results.getLast();
if (rule.getOptimistic() >= worstResult.getConfirmation()) {
return true;
}
return false;
} else {
return true;
}
}
/**
* Test if a rule can be stored in the agenda.
*
* @param rule The rule to consider.
* @return True if the rule can be stored, false otherwise.
*/
private boolean canStoreInNodes(Rule rule) {
if (rule.getObservedNumber() == 0) {
return false;
}
return true;
}
/**
* Test if it is worth calculating the confirmation of a rule.
*
* @param rule The rule to consider.
* @return True if the confirmation can be calculated, false otherwise.
*/
private boolean canCalculateConfirmation(Rule rule) {
if (rule.getObservedFrequency() > m_noiseThreshold) {
return false;
}
return true;
}
/**
* Test if a rule can be added to the results.
*
* @param rule The rule to consider.
* @return True if the rule can be stored, false otherwise.
*/
private boolean canStoreInResults(Rule rule) {
if (rule.getConfirmation() < m_confirmationThreshold) {
return false;
}
if (m_best != 0) {
if (numValuesInResult() < m_best) {
return true;
}
Rule worstResult = (Rule) m_results.getLast();
if (rule.getConfirmation() >= worstResult.getConfirmation()) {
return true;
}
return false;
} else {
return true;
}
}
/**
* Add a rule in the appropriate place in the list of the results,
* according to the confirmation and
* number of counter-instances of the rule. <p>
* Subsumption tests are performed and the rule may not be added. <p>
* Previous results may also be removed because of sumption.
*/
private void addResult(Rule rule) {
Rule current;
boolean added = false;
/* Iterate the list until we find the right place. */
SimpleLinkedList.LinkedListIterator iter = m_results.iterator();
while (iter.hasNext()) {
current = (Rule) iter.next();
if (Rule.confirmationThenObservedComparator.compare(current, rule) > 0) {
iter.addBefore(rule);
added = true;
break;
}
/* Subsumption tests to see if the rule can be added. */
if ((m_subsumption || m_sameClause || m_equivalent)
&& current.subsumes(rule)) {
if (current.numLiterals() == rule.numLiterals()) {
if (current.equivalentTo(rule)) {
/* Equivalent rules. */
if (m_equivalent) {
return;
}
} else {
/* Same clauses. */
if (m_sameClause
&& Rule.confirmationComparator.compare(current, rule) < 0) {
return;
}
}
} else {
/* Subsumption. */
if (m_subsumption
&& Rule.observedComparator.compare(current, rule) <= 0) {
return;
}
}
}
}
if (added == false) {
/* The rule must be added in the end of the results. */
m_results.add(rule);
}
/* Iterate the results with a lower confirmation
* to see if some of them must be removed. */
SimpleLinkedList.LinkedListInverseIterator inverse
= m_results.inverseIterator();
while (inverse.hasPrevious()) {
current = (Rule) inverse.previous();
if (Rule.confirmationThenObservedComparator.compare(current, rule) < 0) {
break;
}
if (current != rule && rule.subsumes(current)) {
if (current.numLiterals() == rule.numLiterals()) {
if (!current.equivalentTo(rule)) {
/* Same clauses. */
if (m_sameClause
&& Rule.confirmationComparator.compare(current, rule) > 0) {
inverse.remove();
}
}
} else {
/* Subsumption. */
if (m_subsumption
&& Rule.observedComparator.compare(rule, current) <= 0) {
inverse.remove();
}
}
}
}
/* Remove the rules with the worst confirmation value
* if there are too many results. */
if (m_best != 0 && numValuesInResult() > m_best) {
Rule worstRule = (Rule) m_results.getLast();
inverse = m_results.inverseIterator();
while (inverse.hasPrevious()) {
current = (Rule) inverse.previous();
if (Rule.confirmationComparator.compare(current, worstRule) < 0) {
break;
}
inverse.remove();
}
}
/* Print the new current values. */
printValues();
}
/**
* Returns default capabilities of the classifier.
*
* @return the capabilities of this classifier
*/
public Capabilities getCapabilities() {
Capabilities result = super.getCapabilities();
result.disableAll();
// attributes
result.enable(Capability.NOMINAL_ATTRIBUTES);
result.enable(Capability.MISSING_VALUES);
// class
result.enable(Capability.NOMINAL_CLASS);
result.enable(Capability.MISSING_CLASS_VALUES);
return result;
}
/**
* Method that launches the search to find the rules with the highest
* confirmation.
*
* @param instances The instances to be used for generating the rules.
* @throws Exception if rules can't be built successfully.
*/
public void buildAssociations(Instances instances) throws Exception {
Frame valuesFrame = null; /* Frame to display the current values. */
/* Initialization of the search. */
if (m_parts == null) {
m_instances = new Instances(instances);
} else {
m_instances = new IndividualInstances(new Instances(instances), m_parts);
}
m_results = new SimpleLinkedList();
m_hypotheses = 0;
m_explored = 0;
m_status = NORMAL;
if (m_classIndex == -1)
m_instances.setClassIndex(m_instances.numAttributes()-1);
else if (m_classIndex < m_instances.numAttributes() && m_classIndex >= 0)
m_instances.setClassIndex(m_classIndex);
else
throw new Exception("Invalid class index.");
// can associator handle the data?
getCapabilities().testWithFail(m_instances);
/* Initialization of the window for current values. */
if (m_printValues == WINDOW) {
m_valuesText = new TextField(37);
m_valuesText.setEditable(false);
m_valuesText.setFont(new Font("Monospaced", Font.PLAIN, 12));
Label valuesLabel = new Label("Best and worst current values:");
Button stop = new Button("Stop search");
stop.addActionListener(new ActionListener() {
public void actionPerformed(ActionEvent e) {
/* Signal the interruption to the search. */
m_status = STOP;
}
});
valuesFrame = new Frame("Tertius status");
valuesFrame.setResizable(false);
valuesFrame.add(m_valuesText, BorderLayout.CENTER);
valuesFrame.add(stop, BorderLayout.SOUTH);
valuesFrame.add(valuesLabel, BorderLayout.NORTH);
valuesFrame.pack();
valuesFrame.setVisible(true);
} else if (m_printValues == OUT) {
System.out.println("Best and worst current values:");
}
Date start = new Date();
/* Build the predicates and launch the search. */
m_predicates = buildPredicates();
beginSearch();
Date end = new Date();
if (m_printValues == WINDOW) {
valuesFrame.dispose();
}
m_time = new Date(end.getTime() - start.getTime());
}
/**
* Run the search.
*/
public void run() {
try {
search();
} catch (OutOfMemoryError e) {
/* Garbage collect what can be collected to be able to continue. */
System.gc();
m_status = MEMORY;
}
endSearch();
}
/**
* Begin the search by starting a new thread.
*/
private synchronized void beginSearch() throws Exception {
/* This method must be synchronized to be able to
* call the wait() method. */
Thread search = new Thread(this);
search.start();
try {
/* Wait for the end of the thread. */
wait();
} catch (InterruptedException e) {
/* Signal the interruption to the search. */
m_status = STOP;
}
}
/**
* End the search by notifying to the waiting thread that it is finished.
*/
private synchronized void endSearch() {
/* This method must be synchronized to be able to
* call the notify() method. */
notify();
}
/**
* Search in the space of hypotheses the rules that have the highest
* confirmation.
* The search starts with the empty rule, other rules are generated by
* refinement.
*/
public void search() {
SimpleLinkedList nodes = new SimpleLinkedList(); /* The agenda. */
Rule currentNode;
SimpleLinkedList children;
SimpleLinkedList.LinkedListIterator iter;
Rule child;
boolean negBody = (m_negation == BODY || m_negation == ALL);
boolean negHead = (m_negation == HEAD || m_negation == ALL);
/* Start with the empty rule. */
nodes.add(new Rule(m_repeat, m_numLiterals, negBody, negHead,
m_classification, m_horn));
/* Print the current values. */
printValues();
/* Explore the rules in the agenda. */
while (m_status != STOP && !nodes.isEmpty()) {
currentNode = (Rule) nodes.removeFirst();
if (canRefine(currentNode)) {
children = currentNode.refine(m_predicates);
iter = children.iterator();
/* Calculate the optimistic estimate of the children and
* consider them for adding to the agenda and to the results. */
while (iter.hasNext()) {
m_hypotheses++;
child = (Rule) iter.next();
child.upDate(m_instances);
if (canCalculateOptimistic(child)) {
child.calculateOptimistic();
if (canExplore(child)) {
m_explored++;
if (canStoreInNodes(child)) {
} else {
iter.remove();
}
if (canCalculateConfirmation(child)) {
child.calculateConfirmation();
if (canStoreInResults(child)) {
addResult(child);
}
}
} else {
iter.remove();
}
} else {
iter.remove();
}
}
/* Since the agenda is already sorted it is more efficient
* to sort the children only and then merge. */
children.sort(Rule.optimisticThenObservedComparator);
nodes.merge(children, Rule.optimisticThenObservedComparator);
} else {
/* The agenda being sorted, it is not worth considering the following
* nodes. */
break;
}
}
}
/**
* returns the results
*
* @return the results
*/
public SimpleLinkedList getResults() {
return m_results;
}
/**
* Print the current best and worst values.
*/
private void printValues() {
if (m_printValues == NO) {
return;
} else {
if (m_results.isEmpty()) {
if (m_printValues == OUT) {
System.out.print("0.000000 0.000000 - 0.000000 0.000000");
} else { //m_printValues == WINDOW
m_valuesText.setText("0.000000 0.000000 - 0.000000 0.000000");
}
} else {
Rule best = (Rule) m_results.getFirst();
Rule worst = (Rule) m_results.getLast();
String values = best.valuesToString() + " - " + worst.valuesToString();
if (m_printValues == OUT) {
System.out.print("\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b"
+ "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b");
System.out.print(values);
} else { //m_printValues == WINDOW
m_valuesText.setText(values);
}
}
}
}
/**
* Outputs the best rules found with their confirmation value and number
* of counter-instances.
* Also gives the number of hypotheses considered and explored, and the
* time needed.
*/
public String toString() {
StringBuffer text = new StringBuffer();
SimpleLinkedList.LinkedListIterator iter = m_results.iterator();
int size = m_results.size();
int i = 0;
text.append("\nTertius\n=======\n\n");
while (iter.hasNext()) {
Rule current = (Rule) iter.next();
text.append(Utils.doubleToString((double) i + 1,
(int) (Math.log(size)
/ Math.log(10) + 1),
0)
+ ". ");
text.append("/* ");
if (m_roc) {
text.append(current.rocToString());
} else {
text.append(current.valuesToString());
}
text.append(" */ ");
text.append(current.toString());
text.append("\n");
i++;
}
text.append("\nNumber of hypotheses considered: " + m_hypotheses);
text.append("\nNumber of hypotheses explored: " + m_explored);
if (m_status == MEMORY) {
text.append("\n\nNot enough memory to continue the search");
} else if (m_status == STOP) {
text.append("\n\nSearch interrupted");
}
return text.toString();
}
/**
* Returns the revision string.
*
* @return the revision
*/
public String getRevision() {
return RevisionUtils.extract("$Revision: 5444 $");
}
/**
* Main method.
*
* @param args the commandline parameters
*/
public static void main(String [] args) {
runAssociator(new Tertius(), args);
}
}