keel.Algorithms.Associative_Classification.ClassifierFuzzyFARCHD

Class RuleBase

java.lang.Object

keel.Algorithms.Associative_Classification.ClassifierFuzzyFARCHD.RuleBase

public class RuleBase
extends java.lang.Object

Title: RuleBase

Description: This class contains the representation of a Rule Set

Copyright: Copyright KEEL (c) 2007

Company: KEEL

Since:

JDK1.5

Version:

1.0

Author:

Written by Jesus Alcala (University of Granada) 09/02/2011

Constructor Summary

Constructors

Constructor and Description
RuleBase() Default Constructor.
RuleBase(DataBase dataBase, myDataset train, int K, int typeInference) Builder

Method Summary

Modifier and Type	Method and Description
void	add(Itemset itemset) It adds a rule to the rule base
void	add(Rule rule) It adds a rule to the rule base
void	add(RuleBase ruleBase) It adds the rules of the RuleBase given.
boolean	BETTER(int a, int b) Maximization
void	clear() Removes the rules stored.
RuleBase	clone() Clone
void	evaluate() Function to evaluate the whole rule base by using the training dataset.
void	evaluate(double[] gene, int[] selected) Function to evaluate the selected rules by using the training dataset and the fuzzy functions stored in the gene given.
int	FRM(double[] example) Predicts the class value for a given example, using the rules and type of inference stored on the RuleBase.
int	FRM(double[] example, int[] selected) Predicts the class value for a given example, using the selected rules and type of inference stored on the RuleBase.
Rule	get(int pos) Function to get a rule from the rule base
double	getAccuracy() Function to return the fitness of the rule base
int	getK() Returns the value of the parameter K.
int	getTypeInference() Returns the type of inference used to build the rules.
int	getUncover() Returns the number of examples uncovered by the rules
int	hasClassUncovered(int[] selected) Indentifies how many classes are uncovered with a selection of rules.
boolean	hasUncover() Checks if there are examples uncovered by the rules.
java.lang.String	printString() It prints the whole rulebase
void	reduceRules(int clas) Function to eliminate the rules that are not needed (Redundant, not enough accurate,...) for a given class.
Rule	remove(int pos) It removes the rule stored in the given position
void	saveFile(java.lang.String filename) It stores the rule base in a given file
void	setDefaultRule() Sets the default rule.
int	size() It returns the number of rules in the rule base
void	sort() Function to sort the rule base

Methods inherited from class java.lang.Object

equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

RuleBase

public RuleBase()

Default Constructor.

RuleBase

public RuleBase(DataBase dataBase,
                myDataset train,
                int K,
                int typeInference)

Builder

Parameters:

dataBase - Data Base

train - Training dataset

K - Covered patterns in the second stage are completely eliminated when they have been covered more than K times.

typeInference - Two option: 0) the class of the rule with the best matching; 1) the class with the best matching

Method Detail

BETTER

public boolean BETTER(int a,
                      int b)

Maximization

Parameters:

a - first number

b - second number

Returns:

boolean true if a is greater than b

clone

public RuleBase clone()

Clone

Overrides:

clone in class java.lang.Object

Returns:

A copy of the Rule Base

add

public void add(Rule rule)

It adds a rule to the rule base

Parameters:

rule - Rule to be added

add

public void add(RuleBase ruleBase)

It adds the rules of the RuleBase given.

Parameters:

ruleBase - Rules to be added

add

public void add(Itemset itemset)

It adds a rule to the rule base

Parameters:

itemset - itemset to be added

get

public Rule get(int pos)

Function to get a rule from the rule base

Parameters:

pos - Position in the rule base where the desired rule is stored

Returns:

The desired rule

size

public int size()

It returns the number of rules in the rule base

Returns:

Rule base's size

sort

public void sort()

Function to sort the rule base

remove

public Rule remove(int pos)

It removes the rule stored in the given position

Parameters:

pos - Position where the rule we want to remove is

Returns:

Removed rule

clear

public void clear()

Removes the rules stored.

getTypeInference

public int getTypeInference()

Returns the type of inference used to build the rules.

Returns:

the type of inference used.

getAccuracy

public double getAccuracy()

Function to return the fitness of the rule base

Returns:

Fitness of the rule base

setDefaultRule

public void setDefaultRule()

Sets the default rule. The default rule classifies all the examples to the majority class.

hasUncover

public boolean hasUncover()

Checks if there are examples uncovered by the rules.

Returns:

True if there are some examples, False otherwise.

getUncover

public int getUncover()

Returns the number of examples uncovered by the rules

Returns:

Number of examples uncovered

getK

public int getK()

Returns the value of the parameter K. (The maximum times covered letting an example active.

Returns:

the value of the parameter K.

evaluate

public void evaluate()

Function to evaluate the whole rule base by using the training dataset.

evaluate

public void evaluate(double[] gene,
                     int[] selected)

Function to evaluate the selected rules by using the training dataset and the fuzzy functions stored in the gene given.

Parameters:

gene - Representation where the fuzzy functions needed to evaluate are stored

selected - Selection of rules to be evaluated

FRM

public int FRM(double[] example)

Predicts the class value for a given example, using the rules and type of inference stored on the RuleBase.

Parameters:

example - Example to be predicted.

Returns:

Class predicted.

FRM

public int FRM(double[] example,
               int[] selected)

Predicts the class value for a given example, using the selected rules and type of inference stored on the RuleBase.

Parameters:

example - Example to be predicted.

selected - Selection of rules to be considered for the class prediction.

Returns:

Class predicted.

hasClassUncovered

public int hasClassUncovered(int[] selected)

Indentifies how many classes are uncovered with a selection of rules.

Parameters:

selected - rules selected to be tested

Returns:

number of classes uncovered.

reduceRules

public void reduceRules(int clas)

Function to eliminate the rules that are not needed (Redundant, not enough accurate,...) for a given class.

Parameters:

clas - class whose rules are being tested

printString

public java.lang.String printString()

It prints the whole rulebase

Returns:

The whole rulebase

saveFile

public void saveFile(java.lang.String filename)

It stores the rule base in a given file

Parameters:

filename - Name for the rulebase file