keel.Algorithms.Rule_Learning.Ripper

Class Ruleset

java.lang.Object

keel.Algorithms.Rule_Learning.Ripper.Ruleset

public class Ruleset
extends java.lang.Object

Representation of a disjuction of rules with a common consecuent. It may be represented as: if (rule1 || rule2) then output=consecuent

Since:

JDK1.2

Version:

1.1

Author:

Written by Alberto Fernández (University of Granada) 01/07/2008, Modified by Xavi Solé (La Salle, Ramón Llull University - Barcelona) 03/12/2008

Constructor Summary

Constructors

Constructor and Description
Ruleset() Constructs an empty ruleset.

Method Summary

Modifier and Type	Method and Description
void	addRule(Rule r) Adds a new rule to the ruleset.
keel.Algorithms.Rule_Learning.Ripper.Stats	apply(MyDataset data) It returns the number of true positives,true negatives,false positives and false negatives of the whole ruleset in a given dataset.
keel.Algorithms.Rule_Learning.Ripper.Stats	apply(MyDataset data, Mask positives, Mask negatives) It returns the number of true positives,true negatives,false positives and false negatives of the whole ruleset in a given dataset.
double	getExceptionCost(MyDataset data) Returns the exception cost for the Minimum Data Length of a dataset given a theory (this ruleset).
double	getExceptionCost(MyDataset data, Mask positives, Mask negatives) Returns the exception cost for the Minimum Data Length of a dataset given a theory (this ruleset).
double	getExceptionCost(MyDataset data, Mask positives, Mask negatives, keel.Algorithms.Rule_Learning.Ripper.IncrementalMask rulesetMask) Returns the exception cost for the Minimum Data Length of a dataset given a theory (this ruleset).
double	getMDL(MyDataset data) Returns the Minimum Data Length of a dataset given a theory (this ruleset).
double	getMDL(MyDataset data, Mask positives, Mask negatives) Returns the Minimum Data Length of a dataset given a theory (this ruleset).
double	getMDL(MyDataset data, Mask positives, Mask negatives, keel.Algorithms.Rule_Learning.Ripper.IncrementalMask rulesetMask) Returns the Minimum Data Length of a dataset given a theory (this ruleset).
Rule	getRule(int pos) Returns the rule in the i-th position of the ruleset.
keel.Algorithms.Rule_Learning.Ripper.IncrementalMask	getRulesetMask(MyDataset data) Returns the combine mask of all the rules in the set.
double	getTheoryCost(MyDataset data) The description length of the theory for the ruleset.
java.lang.String	getType() Returns the common output (consecuent) of the rules in the ruleset.
void	insertRule(Rule r, int pos) Inserts a new rule in a given position of the ruleset.
void	pulish(MyDataset data, Mask positives, Mask negatives) Remove the rules that increase the DL value of the set.
void	removeDuplicates() Removes the duplicated rules
void	removeRule(int pos) Deletes a given rule of the ruleset.
void	setType(java.lang.String type) Sets the common output (consecuent) of the rules in the ruleset.
int	size() Returns the size (number of rules) of the ruleset.
java.lang.String	toString() Returns a string representation of this Ruleset, containing the String representation of each Rule.

Methods inherited from class java.lang.Object

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

Constructor Detail

Ruleset

public Ruleset()

Constructs an empty ruleset.

Method Detail

addRule

public void addRule(Rule r)

Adds a new rule to the ruleset.

Parameters:

r - Rule the new rule

apply

public keel.Algorithms.Rule_Learning.Ripper.Stats apply(MyDataset data)

It returns the number of true positives,true negatives,false positives and false negatives of the whole ruleset in a given dataset. This methods takes into account the right part (consecuent) of the rules

Parameters:

data - MyDataset the dataset

Returns:

number of true positives, false positives, true negatives and false negatives of the whole ruleset in the following order: {tp,tn,fp,fn}

apply

public keel.Algorithms.Rule_Learning.Ripper.Stats apply(MyDataset data,
                                                        Mask positives,
                                                        Mask negatives)

It returns the number of true positives,true negatives,false positives and false negatives of the whole ruleset in a given dataset. (This methods doesn't take into account the right part (consecuent) of the rules).

Parameters:

data - MyDataset the dataset

positives - active positive instances of data

negatives - active negative instances of data

Returns:

number of true positives, false positives, true negatives and false negatives of the whole ruleset in the following order: {tp,tn,fp,fn}

getExceptionCost

public double getExceptionCost(MyDataset data,
                               Mask positives,
                               Mask negatives)

Returns the exception cost for the Minimum Data Length of a dataset given a theory (this ruleset). See [Quinlan95]

Parameters:

data - MyDataset the datasets

positives - Mask active positive entries of data

negatives - Mask active negative entries of data

Returns:

the MDL of data given this ruleset.

getMDL

public double getMDL(MyDataset data,
                     Mask positives,
                     Mask negatives)

Returns the Minimum Data Length of a dataset given a theory (this ruleset). See [Quinlan95]

Parameters:

data - MyDataset the datasets

positives - Mask active positive entries of data

negatives - Mask active negative entries of data

Returns:

the MDL of data given this ruleset.

getExceptionCost

public double getExceptionCost(MyDataset data)

Returns the exception cost for the Minimum Data Length of a dataset given a theory (this ruleset). See [Quinlan95]

Parameters:

data - MyDataset the datasets

Returns:

the MDL of data given this ruleset.

getMDL

public double getMDL(MyDataset data)

Returns the Minimum Data Length of a dataset given a theory (this ruleset). See [Quinlan95]

Parameters:

data - MyDataset the datasets

Returns:

the MDL of data given this ruleset.

getExceptionCost

public double getExceptionCost(MyDataset data,
                               Mask positives,
                               Mask negatives,
                               keel.Algorithms.Rule_Learning.Ripper.IncrementalMask rulesetMask)

Returns the exception cost for the Minimum Data Length of a dataset given a theory (this ruleset). See [Quinlan95]

Parameters:

data - MyDataset the datasets

positives - Mask active positive entries of data

negatives - Mask active negative entries of data

rulesetMask - the combine mask of all rules in the ruleset.

Returns:

the MDL of data given this ruleset.

getMDL

public double getMDL(MyDataset data,
                     Mask positives,
                     Mask negatives,
                     keel.Algorithms.Rule_Learning.Ripper.IncrementalMask rulesetMask)

Returns the Minimum Data Length of a dataset given a theory (this ruleset). See [Quinlan95]

Parameters:

data - MyDataset the datasets

positives - Mask active positive entries of data

negatives - Mask active negative entries of data

rulesetMask - the combine mask of all rules in the ruleset.

Returns:

the MDL of data given this ruleset.

getTheoryCost

public double getTheoryCost(MyDataset data)

The description length of the theory for the ruleset. Computed as the addition of the theory cost for each rule:
0.5* [||k||+ S(t, k, k/t)]
where k is the number of antecedents of the rule; t is the total possible antecedents that could appear in a rule; ||K|| is the universal prior for k , log2*(k) and S(t,k,p) = -k*log2(p)-(n-k)log2(1-p) is the subset encoding length.

Parameters:

data - MyDataset the dataset

Returns:

the description length of the theory for the ruleset

getRule

public Rule getRule(int pos)

Returns the rule in the i-th position of the ruleset.

Parameters:

pos - int position of the rule in the ruleset

Returns:

the rule in the pos-th position of the ruleset.

getType

public java.lang.String getType()

Returns the common output (consecuent) of the rules in the ruleset.

Returns:

the common output (consecuent) of the rules in the ruleset.

getRulesetMask

public keel.Algorithms.Rule_Learning.Ripper.IncrementalMask getRulesetMask(MyDataset data)

Returns the combine mask of all the rules in the set.

Parameters:

data - the dataset

Returns:

the combine mask of all the rules in the set.

insertRule

public void insertRule(Rule r,
                       int pos)

Inserts a new rule in a given position of the ruleset.

Parameters:

r - Rule the new rule

pos - int the position where r must be inserted

removeRule

public void removeRule(int pos)

Deletes a given rule of the ruleset.

Parameters:

pos - int position of the rule in the ruleset.

setType

public void setType(java.lang.String type)

Sets the common output (consecuent) of the rules in the ruleset.

Parameters:

type - String the common output (consecuent) of the rules in the ruleset.

removeDuplicates

public void removeDuplicates()

Removes the duplicated rules

pulish

public void pulish(MyDataset data,
                   Mask positives,
                   Mask negatives)

Remove the rules that increase the DL value of the set.

Parameters:

data - the dataset

positives - the positives exemples

negatives - the negatives exemples

size

public int size()

Returns the size (number of rules) of the ruleset.

Returns:

the size (number of rules) of the ruleset.

toString

public java.lang.String toString()

Returns a string representation of this Ruleset, containing the String representation of each Rule.

Overrides:

toString in class java.lang.Object

Returns:

a string representation of this Ruleset, containing the String representation of each Rule.