at.tuwien.ifs.somtoolbox.data
Class DataBaseSOMLibSparseInputData

java.lang.Object
  at.tuwien.ifs.somtoolbox.data.AbstractSOMLibSparseInputData
      at.tuwien.ifs.somtoolbox.data.DataBaseSOMLibSparseInputData

All Implemented Interfaces:

InputData

public class DataBaseSOMLibSparseInputData
extends AbstractSOMLibSparseInputData

Implements InputData by reading the vector information from a database.

Version:

$Id: DataBaseSOMLibSparseInputData.java 3883 2010-11-02 17:13:23Z frank $

Author:

liegl, Rudolf Mayer

Field Summary

private MySQLConnector

dbConnector

Fields inherited from class at.tuwien.ifs.somtoolbox.data.AbstractSOMLibSparseInputData

classInfo, contentType, dataNames, dim, ERROR_MESSAGE_FILE_FORMAT_CORRUPT, featureMatrixCols, featureMatrixRows, isNormalized, meanVector, mqe0, nameCache, numVectors, rand, source, templateVector

Fields inherited from interface at.tuwien.ifs.somtoolbox.data.InputData

inputFileNameSuffix, MISSING_VALUE

Constructor Summary

DataBaseSOMLibSparseInputData(MySQLConnector dbConnector, boolean sparse, boolean norm, int numCacheBlocks, long seed)

DataBaseSOMLibSparseInputData(MySQLConnector dbConnector, boolean sparse, String classInfoFileName, boolean norm, int numCacheBlocks, long seed)

Method Summary

protected void	calculateMeanVector() Initialise the mean-vector.
private String	getErrorMessage(Exception e)
InputDatum	getInputDatum(int index) Get an input datum with a specified index.
double[]	getInputVector(int d) Get the vector for the input datum of the specified index
double	getValue(int x, int y) Returns the value of the y-th feature of input vector x.
double	mqe0(DistanceMetric metric) Calculates the mean quantisation error of the top-level unit.
private cern.colt.matrix.DoubleMatrix1D	readVectorFromDataBase(int rowindex) Read a vector from the database, identified by the given index.
private void	readVectorMetaDataFromDataBase() Read input meta data from the database.
InputData	subset(String[] names) TODO: this reads the whole subset into the memory.

Methods inherited from class at.tuwien.ifs.somtoolbox.data.AbstractSOMLibSparseInputData

classInformation, create, dim, equals, getByNameDistanceSorted, getContentType, getData, getData, getDataIntervals, getDataSource, getDistanceMatrix, getDistances, getFeatureDensities, getFeatureMatrixColumns, getFeatureMatrixRows, getFileNameSuffix, getFormatName, getInputDatum, getInputDatum, getInputDatumIndex, getLabel, getLabels, getMaxValue, getMeanVector, getMeanVector, getMinValue, getNearestDistances, getNearestN, getNearestN, getNearestNUnsorted, getRandomInputDatum, initDistanceMatrix, isNormalizedToUnitLength, numVectors, setClassInfo, setTemplateVector, templateVector, transformValues

Methods inherited from class java.lang.Object

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

Field Detail

dbConnector

private MySQLConnector dbConnector

Constructor Detail

DataBaseSOMLibSparseInputData

public DataBaseSOMLibSparseInputData(MySQLConnector dbConnector,
                                     boolean sparse,
                                     boolean norm,
                                     int numCacheBlocks,
                                     long seed)

DataBaseSOMLibSparseInputData

public DataBaseSOMLibSparseInputData(MySQLConnector dbConnector,
                                     boolean sparse,
                                     String classInfoFileName,
                                     boolean norm,
                                     int numCacheBlocks,
                                     long seed)
                              throws SOMToolboxException

Throws:

SOMToolboxException

Method Detail

readVectorMetaDataFromDataBase

private void readVectorMetaDataFromDataBase()

Read input meta data from the database. This includes:

• Number of input vectors (numVectors())
• Vector dimension (dim())
• The input data labels / names (dataNames)
• Initialising the name cache (nameCache)
• Calculating the mean vector (getMeanVector())

getInputDatum

public InputDatum getInputDatum(int index)

Description copied from interface: InputData

Get an input datum with a specified index.

Parameters:

index - the index of the input datum.

Returns:

the input datum.

getInputVector

public double[] getInputVector(int d)

Description copied from interface: InputData

Get the vector for the input datum of the specified index

readVectorFromDataBase

private cern.colt.matrix.DoubleMatrix1D readVectorFromDataBase(int rowindex)

Read a vector from the database, identified by the given index.

Parameters:

rowindex - the index of the vector.

Returns:

the input vector.

mqe0

public double mqe0(DistanceMetric metric)

Description copied from interface: InputData

Calculates the mean quantisation error of the top-level unit.

Parameters:

metric - the metric to use for distance calculation.

Returns:

the mqe0.

subset

public InputData subset(String[] names)

TODO: this reads the whole subset into the memory.

Parameters:

names - the label names of the desired subset data.

Returns:

a subset of the data.

See Also:

InputData.subset(java.lang.String[])

calculateMeanVector

protected void calculateMeanVector()
                            throws SQLException

Initialise the mean-vector.

Throws:

SQLException

getErrorMessage

private String getErrorMessage(Exception e)

getValue

public double getValue(int x,
                       int y)

Description copied from interface: InputData

Returns the value of the y-th feature of input vector x.