|
||||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | |||||||||
SUMMARY: NESTED | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |
java.lang.Object | +--weka.roughset.RelativeReductComputation
Class for Relative Reducts, i.e. sets of attributes that maintain the discernability function between instances.
Constructor Summary | |
RelativeReductComputation(Instances i)
|
|
RelativeReductComputation(Instances i,
int d)
|
Method Summary | |
Reducts |
dynamicReducts(int choices,
double confidence,
double maxError,
double stability,
int numReducts)
Obtains a set of reducts by using the method described in Dynamic reducts as a tool for extracting laws from decisions tables, Jan G. |
int |
getDebug()
Get the value of debug. |
static void |
main(String[] args)
|
Reduct |
minimalRelativeReduct(int index)
Returns the minimal set of attributes that discerns instance at position index from other instances. |
Reduct |
randomReduct()
Works as semiMinimalReduct , only we use a
non-deterministic algorithm that uses
weka.roughset.SortedClasses.randomBestAttribute(Attributes
attrs) for selecting each successive attribute. |
static Reduct |
semiMinimalReduct(Attributes old,
Instances instances)
Starting from the attributes in old, build a superreduct that maintains the indiscernibility relation that is given by the whole set of attributes by using a greedy algorithm that selects the attribute that discerns the most objects as returned by the weka.roughset.SortedClasses.bestAttribute(Attributes
attrs) method. |
void |
setDebug(int v)
Set the value of debug. |
Reducts |
someProperReducts(int numIterations)
This simple heuristic takes a random reduct and reduces it into a set of proper reducts |
static Reducts |
someProperReducts(Reduct superReduct,
int numIterations)
|
Reduct |
wroblewskiReduct(int[] perm)
Obtaina a reduct by using the method described in Genetic algorithms in decomposition of classification problems, Jakub Wróblewski in Rough Sets in Knowledge Discovery 2: Applications, Case Studies and Software Systems, 1998 |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
Constructor Detail |
public RelativeReductComputation(Instances i)
public RelativeReductComputation(Instances i, int d)
Method Detail |
public int getDebug()
public void setDebug(int v)
v
- Value to assign to debug.public Reduct minimalRelativeReduct(int index)
index
- an int
value
Reduct
valuepublic static Reduct semiMinimalReduct(Attributes old, Instances instances)
weka.roughset.SortedClasses.bestAttribute(Attributes
attrs)
method.
old
- an Attributes
valueinstances
- an Instances
value
Reduct
valuepublic Reduct randomReduct()
semiMinimalReduct
, only we use a
non-deterministic algorithm that uses
weka.roughset.SortedClasses.randomBestAttribute(Attributes
attrs)
for selecting each successive attribute.
Reduct
valuepublic Reduct wroblewskiReduct(int[] perm)
perm
- an int[]
value
Reduct
valuepublic Reducts dynamicReducts(int choices, double confidence, double maxError, double stability, int numReducts) throws Exception
choices
- an int
valueconfidence
- a double
valuemaxError
- a double
valuestability
- a double
valuenumReducts
- an int
value
Reducts
value
Exception
- if an error occurspublic static Reducts someProperReducts(Reduct superReduct, int numIterations) throws Exception
Exception
public Reducts someProperReducts(int numIterations) throws Exception
Exception
public static void main(String[] args)
|
||||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | |||||||||
SUMMARY: NESTED | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |