keel.Algorithms.Rule_Learning.CN2

Class CN2

java.lang.Object

keel.Algorithms.Rule_Learning.CN2.CN2

public class CN2
extends java.lang.Object

Title: Main class of the algorithm

Description: It contains the esential methods for the CN2 algorithm

Created: November 26th 2004

Copyright: Copyright (c) 2004

Company: KEEL

Since:

JDK1.5

Version:

1.6

Author:

Alberto Fernández (University of Granada) 26/11/2004

Constructor Summary

Constructors

Constructor and Description
CN2() Default builder
CN2(java.lang.String ftrain, java.lang.String feval, java.lang.String ftest, java.lang.String foutputTr, java.lang.String foutputTst, java.lang.String fsal, int starSize, double _covered, int _accuracy) CN2 class builder It does a local copy of the filenames for their posterior use.

Method Summary

Modifier and Type	Method and Description
boolean	everythingOK() It checks if some of the preconditions are not satisfied: There are any continuous value or there was a problem while reading the data files
void	execute() We execute here the CN2 algorithm and we create the necessary output data

Methods inherited from class java.lang.Object

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

Constructor Detail

CN2

public CN2()

Default builder

CN2

public CN2(java.lang.String ftrain,
           java.lang.String feval,
           java.lang.String ftest,
           java.lang.String foutputTr,
           java.lang.String foutputTst,
           java.lang.String fsal,
           int starSize,
           double _covered,
           int _accuracy)

CN2 class builder It does a local copy of the filenames for their posterior use.
Then, it obtains all data from file and stores it in a format recognizable for the program.
Finally, it creates all possible selectors for the dataset and stores them.

Parameters:

ftrain - Name of the input training file

feval - Name of the input validation file

ftest - Name of the input test file

foutputTr - Name of the output training file

foutputTst - Name of the output test file

fsal - Name of the output information file

starSize - It is the maximum size for the star in the search process (beam search)

_covered - The percentage of maximum examples to cover

_accuracy - It refers wether the complete selectors will be employed (disjunctions).

Method Detail

everythingOK

public boolean everythingOK()

It checks if some of the preconditions are not satisfied: There are any continuous value or there was a problem while reading the data files

Returns:

boolean true if the algorithm can run normally, false in other case

execute

public void execute()

We execute here the CN2 algorithm and we create the necessary output data