Package ai.libs.jaicore.ml.core.filter.sampling.inmemory.stratified.sampling

Class DiscretizationHelper

java.lang.Object

ai.libs.jaicore.ml.core.filter.sampling.inmemory.stratified.sampling.DiscretizationHelper

Type Parameters:

I - The instance type

All Implemented Interfaces:

org.api4.java.common.control.ILoggingCustomizable

public class DiscretizationHelper
extends java.lang.Object
implements org.api4.java.common.control.ILoggingCustomizable

This helper class provides methods that are required in order to discretize numeric attributes.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	DiscretizationHelper.DiscretizationStrategy

Constructor Summary

Constructors

Constructor	Description
DiscretizationHelper()

Method Summary

Modifier and Type	Method	Description
java.util.Map<java.lang.Integer,AttributeDiscretizationPolicy>	createDefaultDiscretizationPolicies(org.api4.java.ai.ml.core.dataset.IDataset<?> dataset, java.util.List<java.lang.Integer> indices, java.util.Map<java.lang.Integer,java.util.Set<java.lang.Object>> attributeValues, DiscretizationHelper.DiscretizationStrategy discretizationStrategy, int numberOfCategories)	This method creates a default discretization policy for each numeric attribute in the attributes that have to be considered for stratum assignment.
protected int	discretize(double value, AttributeDiscretizationPolicy policy)	Discretizes the particular provided value.
protected void	discretizeAttributeValues(java.util.Map<java.lang.Integer,AttributeDiscretizationPolicy> discretizationPolicies, java.util.Map<java.lang.Integer,java.util.Set<java.lang.Object>> attributeValues)	Discretizes the given attribute values with respect to the provided policies
AttributeDiscretizationPolicy	equalLengthPolicy(java.util.List<java.lang.Double> numericValues, int numberOfCategories)	Creates an equal length policy for the given values with respect to the given number of categories.
AttributeDiscretizationPolicy	equalSizePolicy(java.util.List<java.lang.Double> numericValues, int numberOfCategories)	Creates an equal size policy for the given values with respect to the given number of categories.
java.lang.String	getLoggerName()
void	setLoggerName(java.lang.String name)

Methods inherited from class java.lang.Object

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

Constructor Detail

DiscretizationHelper

public DiscretizationHelper()

Method Detail

createDefaultDiscretizationPolicies

public java.util.Map<java.lang.Integer,AttributeDiscretizationPolicy> createDefaultDiscretizationPolicies(org.api4.java.ai.ml.core.dataset.IDataset<?> dataset,
                                                                                                                java.util.List<java.lang.Integer> indices,
                                                                                                                java.util.Map<java.lang.Integer,java.util.Set<java.lang.Object>> attributeValues,
                                                                                                                DiscretizationHelper.DiscretizationStrategy discretizationStrategy,
                                                                                                                int numberOfCategories)

This method creates a default discretization policy for each numeric attribute in the attributes that have to be considered for stratum assignment.

Parameters:

dataset - The data set that has to be sampled

indices - Indices of the attributes that have to be considered for stratum assignment

attributeValues - Values of the relevant attributes

discretizationStrategy - The discretization strategy that has to be used

numberOfCategories - The number of categories to which the numeric values have to be assigned

Returns:

equalSizePolicy

public AttributeDiscretizationPolicy equalSizePolicy(java.util.List<java.lang.Double> numericValues,
                                                     int numberOfCategories)

Creates an equal size policy for the given values with respect to the given number of categories. An equal size policy is a policy where the length of the intervals is chosen such that in each interval there are equally many values.

Parameters:

numericValues - Distinct attribute values in ascending order

numberOfCategories - Number of categories

Returns:

The created discretization policy consisting of one interval per category

equalLengthPolicy

public AttributeDiscretizationPolicy equalLengthPolicy(java.util.List<java.lang.Double> numericValues,
                                                       int numberOfCategories)

Creates an equal length policy for the given values with respect to the given number of categories. An equal length policy is a policy where the length of the intervals is the same for all intervals.

Parameters:

numericValues - Distinct attribute values in ascending order

numberOfCategories - Number of categories

Returns:

The created discretization policy consisting of one interval per category

discretizeAttributeValues

protected void discretizeAttributeValues(java.util.Map<java.lang.Integer,AttributeDiscretizationPolicy> discretizationPolicies,
                                         java.util.Map<java.lang.Integer,java.util.Set<java.lang.Object>> attributeValues)

Discretizes the given attribute values with respect to the provided policies

Parameters:

discretizationPolicies -

attributeValues -

discretize

protected int discretize(double value,
                         AttributeDiscretizationPolicy policy)

Discretizes the particular provided value. Discretization in this case means to replace the original value by a categorical value. The categorical value is simply the index of the interval the value was assigned to.

Parameters:

value - The (numeric) value to be discretized

policy - The policy that has to be used for discretization

Returns:

getLoggerName

public java.lang.String getLoggerName()

Specified by:

getLoggerName in interface org.api4.java.common.control.ILoggingCustomizable

setLoggerName

public void setLoggerName(java.lang.String name)

Specified by:

setLoggerName in interface org.api4.java.common.control.ILoggingCustomizable