meka.core

Class F

java.lang.Object

meka.core.F

public abstract class F
extends Object

F.java - transform/Filter Operations on instances. Transform 'D' and 'x' in many different ways wrt the labels. Note that there are similar/related functions in MLUtils.java. In the future these should be combined together.

Author:

Jesse Read

See Also:

MLUtils

Constructor Summary

Constructors

Constructor and Description
F()

Method Summary

Modifier and Type	Method and Description
static weka.core.Instances	keepLabels(weka.core.Instances D, int L, int[] j) Remove Indices - Remove some labels (assume they are the first L attributes) from D.
static weka.core.Instance	meka2mulan(weka.core.Instance x, int L) meka2mulan - Move L label attributes from the beginning to end of attribute space of an Instance.
static weka.core.Instances	meka2mulan(weka.core.Instances D, int L) meka2mulan - Move L label attributes from the beginning to end of attribute space of an Instances.
static weka.core.Instances	mulan2meka(weka.core.Instances D, int L) mulan2meka - Move label attributes from the End to the Beginning of attribute space (MULAN format to MEKA format).
static weka.core.Instances	remove(weka.core.Instances D, int[] indices, boolean inv) Remove Indices - Remove attribute indices 'indices' from 'D'.
static weka.core.Instances	removeLabels(weka.core.Instances D, int L) Remove Indices - Remove ALL labels (assume they are the first L attributes) from D.
static void	reorderLabels(weka.core.Instances D, int[] s) ReorderLabels - swap values of y[1] to y[L] according to s[].

Methods inherited from class java.lang.Object

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

Constructor Detail

F

public F()

Method Detail

meka2mulan

public static final weka.core.Instances meka2mulan(weka.core.Instances D,
                                                   int L)

meka2mulan - Move L label attributes from the beginning to end of attribute space of an Instances. Necessary because MULAN assumes label attributes are at the end, not the beginning. (the extra time for this process is not counted in the running-time analysis of published work).

meka2mulan

public static final weka.core.Instance meka2mulan(weka.core.Instance x,
                                                  int L)

meka2mulan - Move L label attributes from the beginning to end of attribute space of an Instance. Necessary because MULAN assumes label attributes are at the end, not the beginning. (the extra time for this process is not counted in the running-time analysis of published work).

mulan2meka

public static final weka.core.Instances mulan2meka(weka.core.Instances D,
                                                   int L)

mulan2meka - Move label attributes from the End to the Beginning of attribute space (MULAN format to MEKA format). Note: can use e.g.: java weka.filters.unsupervised.attribute.Reorder -i thyroid.arff -R 30-last,1-29" See also: F.reorderLabels(D,s)

reorderLabels

public static void reorderLabels(weka.core.Instances D,
                                 int[] s)
                          throws Exception

ReorderLabels - swap values of y[1] to y[L] according to s[].

Parameters:

s - new indices order (supposing that it contains the first s.length indices)

Throws:

Exception

remove

public static weka.core.Instances remove(weka.core.Instances D,
                                         int[] indices,
                                         boolean inv)
                                  throws Exception

Remove Indices - Remove attribute indices 'indices' from 'D'.

Parameters:

D - Dataset

indices - attribute indices to remove/keep

inv - if true, then keep 'indices'

Returns:

New dataset with 'indices' removed.

Throws:

Exception

removeLabels

public static weka.core.Instances removeLabels(weka.core.Instances D,
                                               int L)
                                        throws Exception

Remove Indices - Remove ALL labels (assume they are the first L attributes) from D.

Parameters:

D - Dataset

L - number of labels

Returns:

New dataset with labels removed.

Throws:

Exception

keepLabels

public static weka.core.Instances keepLabels(weka.core.Instances D,
                                             int L,
                                             int[] j)
                                      throws Exception

Remove Indices - Remove some labels (assume they are the first L attributes) from D.

Parameters:

D - Dataset

L - number of labels

j - indices of labels to keep

Returns:

New dataset with labels removed.

Throws:

Exception