org.datavec.api.transform.transform.sequence

Class SequenceMovingWindowReduceTransform

java.lang.Object

org.datavec.api.transform.transform.sequence.SequenceMovingWindowReduceTransform

All Implemented Interfaces:

Serializable, ColumnOp, Transform

public class SequenceMovingWindowReduceTransform
extends Object
implements Transform

SequenceMovingWindowReduceTransform Adds a new column, where the value is derived by:
(a) using a window of the last N values in a single column,
(b) Apply a reduction op on the window to calculate a new value
for example, this transformer can be used to implement a simple moving average of the last N values, or determine the minimum or maximum values in the last N time steps. For a simple moving average, length 20: new SequenceMovingWindowReduceTransform("myCol", 20, ReduceOp.Mean)

Author:

Alex Black

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SequenceMovingWindowReduceTransform.EdgeCaseHandling Enumeration to specify how each cases are handled: For example, for a look back period of 20, how should the first 19 output values be calculated? Default: Perform your former reduction as normal, with as many values are available SpecifiedValue: use the given/specified value instead of the actual output value.

Constructor Summary

Constructors

Constructor and Description
SequenceMovingWindowReduceTransform(String columnName, int lookback, ReduceOp op)
SequenceMovingWindowReduceTransform(String columnName, String newColumnName, int lookback, ReduceOp op, SequenceMovingWindowReduceTransform.EdgeCaseHandling edgeCaseHandling, Writable edgeCaseValue)

Method Summary

Modifier and Type	Method and Description
String	columnName() Returns a singular column name this op is meant to run on
String[]	columnNames() Returns column names this op is meant to run on
static String	defaultOutputColumnName(String originalName, int lookback, ReduceOp op)
Schema	getInputSchema() Getter for input schema
List<Writable>	map(List<Writable> writables) Transform a writable in to another writable
Object	map(Object input) Transform an object in to another object
List<List<Writable>>	mapSequence(List<List<Writable>> sequence) Transform a sequence
Object	mapSequence(Object sequence) Transform a sequence
String	outputColumnName() The output column name after the operation has been applied
String[]	outputColumnNames() The output column names This will often be the same as the input
void	setInputSchema(Schema inputSchema) Set the input schema.
String	toString()
Schema	transform(Schema inputSchema) Get the output schema for this transformation, given an input schema

Methods inherited from class java.lang.Object

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

Constructor Detail

SequenceMovingWindowReduceTransform

public SequenceMovingWindowReduceTransform(String columnName,
                                           int lookback,
                                           ReduceOp op)

Parameters:

columnName - Column name to perform windowing on

lookback - Look back period for windowing

op - Reduction operation to perform on each window

SequenceMovingWindowReduceTransform

public SequenceMovingWindowReduceTransform(String columnName,
                                           String newColumnName,
                                           int lookback,
                                           ReduceOp op,
                                           SequenceMovingWindowReduceTransform.EdgeCaseHandling edgeCaseHandling,
                                           Writable edgeCaseValue)

Parameters:

columnName - Column name to perform windowing on

newColumnName - Name of the new output column (with results)

lookback - Look back period for windowing

op - Reduction operation to perform on each window

edgeCaseHandling - How the 1st steps should be handled (positions in sequence with indices less then the look-back period)

edgeCaseValue - Used only with EdgeCaseHandling.SpecifiedValue, maybe null otherwise

Method Detail

defaultOutputColumnName

public static String defaultOutputColumnName(String originalName,
                                             int lookback,
                                             ReduceOp op)

transform

public Schema transform(Schema inputSchema)

Description copied from interface: ColumnOp

Get the output schema for this transformation, given an input schema

Specified by:

transform in interface ColumnOp

setInputSchema

public void setInputSchema(Schema inputSchema)

Description copied from interface: ColumnOp

Set the input schema.

Specified by:

setInputSchema in interface ColumnOp

getInputSchema

public Schema getInputSchema()

Description copied from interface: ColumnOp

Getter for input schema

Specified by:

getInputSchema in interface ColumnOp

Returns:

map

public List<Writable> map(List<Writable> writables)

Description copied from interface: Transform

Transform a writable in to another writable

Specified by:

map in interface Transform

Parameters:

writables - the record to transform

Returns:

the transformed writable

mapSequence

public List<List<Writable>> mapSequence(List<List<Writable>> sequence)

Description copied from interface: Transform

Transform a sequence

Specified by:

mapSequence in interface Transform

map

public Object map(Object input)

Transform an object in to another object

Specified by:

map in interface Transform

Parameters:

input - the record to transform

Returns:

the transformed writable

mapSequence

public Object mapSequence(Object sequence)

Transform a sequence

Specified by:

mapSequence in interface Transform

Parameters:

sequence -

toString

public String toString()

Overrides:

toString in class Object

outputColumnName

public String outputColumnName()

The output column name after the operation has been applied

Specified by:

outputColumnName in interface ColumnOp

Returns:

the output column name

outputColumnNames

public String[] outputColumnNames()

The output column names This will often be the same as the input

Specified by:

outputColumnNames in interface ColumnOp

Returns:

the output column names

columnNames

public String[] columnNames()

Returns column names this op is meant to run on

Specified by:

columnNames in interface ColumnOp

Returns:

columnName

public String columnName()

Returns a singular column name this op is meant to run on

Specified by:

columnName in interface ColumnOp

Returns: