org.encog.ml.data.buffer.codec

Class CSVDataCODEC

java.lang.Object

org.encog.ml.data.buffer.codec.CSVDataCODEC

All Implemented Interfaces:

DataSetCODEC

public class CSVDataCODEC
extends java.lang.Object
implements DataSetCODEC

A CODEC used to read/write data from/to a CSV data file. There are two constructors provided, one is for reading, the other for writing. Make sure you use the correct one for your intended purpose. This CODEC is typically used with the BinaryDataLoader, to load external data into the Encog binary training format.

Constructor Summary

Constructors

Constructor and Description
CSVDataCODEC(java.io.File theFile, CSVFormat theFormat, boolean theExpectSignificance) Constructor to create CSV from binary..
CSVDataCODEC(java.io.File theFile, CSVFormat theFormat, boolean theHeaders, int theInputCount, int theIdealCount, boolean theExpectSignificance) Create a CODEC to load data from CSV to binary.

Method Summary

Modifier and Type	Method and Description
void	close() Close any open files.
int	getIdealSize()
int	getInputSize()
void	prepareRead() Prepare to read from the CSV file.
void	prepareWrite(int recordCount, int inputSize, int idealSize) Prepare to write to a CSV file.
boolean	read(double[] input, double[] ideal, double[] significance) Read one record of data from a CSV file.
void	write(double[] input, double[] ideal, double significance) Write one record of data to a CSV file.

Methods inherited from class java.lang.Object

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

Constructor Detail

CSVDataCODEC

public CSVDataCODEC(java.io.File theFile,
                    CSVFormat theFormat,
                    boolean theExpectSignificance)

Constructor to create CSV from binary..

Parameters:

theFile - The CSV file to create.

theFormat - The format for that CSV file.

theExpectSignificance - True, if a significance column is expected.

CSVDataCODEC

public CSVDataCODEC(java.io.File theFile,
                    CSVFormat theFormat,
                    boolean theHeaders,
                    int theInputCount,
                    int theIdealCount,
                    boolean theExpectSignificance)

Create a CODEC to load data from CSV to binary.

Parameters:

theFile - The CSV file to load.

theFormat - The format that the CSV file is in.

theHeaders - True, if there are headers.

theInputCount - The number of input columns.

theIdealCount - The number of ideal columns.

theExpectSignificance - True, if a significance column is expected.

Method Detail

close

public void close()

Close any open files.

Specified by:

close in interface DataSetCODEC

getIdealSize

public int getIdealSize()

Specified by:

getIdealSize in interface DataSetCODEC

Returns:

The size of the ideal data.

getInputSize

public int getInputSize()

Specified by:

getInputSize in interface DataSetCODEC

Returns:

The size of the input data.

prepareRead

public void prepareRead()

Prepare to read from the CSV file.

Specified by:

prepareRead in interface DataSetCODEC

prepareWrite

public void prepareWrite(int recordCount,
                         int inputSize,
                         int idealSize)

Prepare to write to a CSV file.

Specified by:

prepareWrite in interface DataSetCODEC

Parameters:

recordCount - The total record count, that will be written.

inputSize - The input size.

idealSize - The ideal size.

read

public boolean read(double[] input,
                    double[] ideal,
                    double[] significance)

Read one record of data from a CSV file.

Specified by:

read in interface DataSetCODEC

Parameters:

input - The input data array.

ideal - The ideal data array.

significance - The significance. The first element will be modified to hold the significance.

Returns:

True, if there is more data to be read.

write

public void write(double[] input,
                  double[] ideal,
                  double significance)

Write one record of data to a CSV file.

Specified by:

write in interface DataSetCODEC

Parameters:

input - The input data array.

ideal - The ideal data array.

significance - The significance (full = 1.0)