org.encog.ml.data.versatile

Class VersatileMLDataSet

java.lang.Object

org.encog.ml.data.versatile.MatrixMLDataSet

org.encog.ml.data.versatile.VersatileMLDataSet

All Implemented Interfaces:

java.lang.Iterable<MLDataPair>, MLDataSet

public class VersatileMLDataSet
extends MatrixMLDataSet

The versatile dataset supports several advanced features. 1. it can directly read and normalize from a CSV file. 2. It supports virtual time-boxing for time series data (the data is NOT expanded in memory). 3. It can easily be segmented into smaller datasets.

Nested Class Summary

Nested classes/interfaces inherited from class org.encog.ml.data.versatile.MatrixMLDataSet

MatrixMLDataSet.MatrixMLDataSetIterator

Constructor Summary

Constructors

Constructor and Description
VersatileMLDataSet(VersatileDataSource theSource) Construct the data source.

Method Summary

Modifier and Type	Method and Description
void	analyze() Analyze the input and determine max, min, mean, etc.
void	defineInput(ColumnDefinition col) Define an input column.
void	defineMultipleOutputsOthersInput(ColumnDefinition[] outputColumns) Define multiple output columns, all others as inputs.
void	defineOutput(ColumnDefinition col) Define an output column.
void	defineSingleOutputOthersInput(ColumnDefinition outputColumn) Define a single column as an output column, all others as inputs.
ColumnDefinition	defineSourceColumn(java.lang.String name, ColumnType colType) Define a source column.
ColumnDefinition	defineSourceColumn(java.lang.String name, int index, ColumnType colType) Define a source column.
void	divide(java.util.List<DataDivision> dataDivisionList, boolean shuffle, GenerateRandom rnd) Divide, and optionally shuffle, the dataset.
NormalizationHelper	getNormHelper()
void	normalize() Normalize the data set, and allocate memory to hold it.
void	setNormHelper(NormalizationHelper helper)

Methods inherited from class org.encog.ml.data.versatile.MatrixMLDataSet

add, add, add, close, get, getCalculatedIdealSize, getCalculatedInputSize, getData, getIdealSize, getInputSize, getLagWindowSize, getLeadWindowSize, getMask, getRecord, getRecordCount, isSupervised, iterator, openAdditional, setCalculatedIdealSize, setCalculatedInputSize, setData, setLagWindowSize, setLeadWindowSize, size

Methods inherited from class java.lang.Object

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

Methods inherited from interface java.lang.Iterable

forEach, spliterator

Constructor Detail

VersatileMLDataSet

public VersatileMLDataSet(VersatileDataSource theSource)

Construct the data source.

Parameters:

theSource - The data source.

Method Detail

analyze

public void analyze()

Analyze the input and determine max, min, mean, etc.

normalize

public void normalize()

Normalize the data set, and allocate memory to hold it.

defineSourceColumn

public ColumnDefinition defineSourceColumn(java.lang.String name,
                                           int index,
                                           ColumnType colType)

Define a source column. Used when the file does not contain headings.

Parameters:

name - The name of the column.

index - The index of the column.

colType - The column type.

Returns:

The column definition.

getNormHelper

public NormalizationHelper getNormHelper()

Returns:

the helper

setNormHelper

public void setNormHelper(NormalizationHelper helper)

Parameters:

helper - the helper to set

divide

public void divide(java.util.List<DataDivision> dataDivisionList,
                   boolean shuffle,
                   GenerateRandom rnd)

Divide, and optionally shuffle, the dataset.

Parameters:

dataDivisionList - The desired divisions.

shuffle - True, if we should shuffle.

rnd - Random number generator, often with a specific seed.

defineOutput

public void defineOutput(ColumnDefinition col)

Define an output column.

Parameters:

col - The output column.

defineInput

public void defineInput(ColumnDefinition col)

Define an input column.

Parameters:

col - The input column.

defineSingleOutputOthersInput

public void defineSingleOutputOthersInput(ColumnDefinition outputColumn)

Define a single column as an output column, all others as inputs.

Parameters:

outputColumn - The output column.

defineSourceColumn

public ColumnDefinition defineSourceColumn(java.lang.String name,
                                           ColumnType colType)

Define a source column.

Parameters:

name - The name of the source column.

colType - The column type.

Returns:

The column definition.

defineMultipleOutputsOthersInput

public void defineMultipleOutputsOthersInput(ColumnDefinition[] outputColumns)

Define multiple output columns, all others as inputs.

Parameters:

outputColumns - The output columns.