Java SOMToolbox Applications & Tools
This page contains a reference documentation of the applications and tools available in the Java SOMToolbox.
The overview on existing applications is also available when running the somtoolbox.sh
script (somtoolbox.bat
on Windows systems), while the detailed help for each application is available when running somtoolbox.sh
<applicationName> --help
(somtoolbox.bat <applicationName> --help
on Windows systems).
Training Applications | |
GHSOM | The Growing Hierarchical SOM grows a hierarchy of maps, depending on the structure of the data set. |
GrowingSOM | Provides a Growing Grid and a static SOM (with the growth parameters disabled) |
MnemonicSOM | MenmonicSOM has a grid which is not fully occupied with units. |
SOMTrainer | Graphical Interface to train a SOM |
Viewer Applications | |
SOMViewer | An interactive viewer for exploring SOMs, using different visualisations |
UnitFileViewer | Plots a unit file, can be used especially to plot a 3D-SOM |
Utils Applications | |
AttendeeMapper | Writes an HTML output of the map, with highlighting certain data items |
DataMapper | Maps inputs to an already trained SOM. |
DataSetViewer | |
HTMLOutputter | Creates an HTML representation of the Map. |
RhythmPatternsMatrixVisualisationSaver | Save rhythm patterns matrix visualisation of an input file as images to a file. |
SOMMerger | Combines the weight vectors of one or more SOM maps to an input vector file |
SimilarityRetrieval | Performs similarity retrieval in a given database (vector file) |
SimilarityRetrievalGUI | GUI for similarity retrieval |
SimilarityRetrievalWriter | Performs similarity retrieval in a given database (vector file) |
SomFilePacker | Packer to create autonomous SomFiles |
TemplateVectorComparator | Compares the contents of two template vectors |
TrajectoryOutputter | Generates a graphical representation of a trajectory of the given points over the map |
VisualisationImageSaver | Save Visualisations of a map as images to a file. |
Helper Applications | |
ClassSubsetGenerator | Generates a subset and class assignment based on the given class file |
DataWinnerMappingWriter | Writes a data winner mapping file from a trained map |
DatasetRandomiser | Randomises data sets |
DistanceMatrixWriter | Writes a distance matrix for the given data |
InputDataFileFormatConverter | Converts between various file formats for input data. |
LabelSOM | Implements the LabelSOM labelling method |
LagusKeywordLabeler | Implements the LagusKeyword labelling method |
MapFileFormatConverter | Converts between various file formats for trained SOMs. |
MapRotator | Rotates a map by the given degrees, and writes a new unit- and weight-vector file |
MnemonicSOMGenerator | MnemonicSOMGenerator allows to create an arbitrary shaped SOM (unit file) |
QualityMeasureComputer | Wrapper for the individual Quality Measures |
RGBPaletteConverter | Converts palettes given as RGB values to the XML format used in the SOMToolbox. |
SOMLibInputConcatenator | Merges two or more SOMLib Input files, i.e. vector and template files |
SOMLibVectorNormalization | Handles the normalization of vector files in SOMLib format |
SOMLibZeroVectorRemover | Removes zero vectors (i.e. with 0 in all their components) from files in SOMLib format |
StringReplacer | Replaces strings by a replacement in the given file |
VectorFile2DatabaseImporter | Imports input and template vector files to a database |
VectorFileAppender | Append SOMLibVectorFiles |
VectorFileConcatenator | Merge SOMLibVectorFiles |
VectorFileRewriter | Replace labels in SOMLibVectorFiles |
Training Applications
GHSOM
The Growing Hierarchical SOM grows a hierarchy of maps, depending on the structure of the data set.
Usage: java at.tuwien.ifs.somtoolbox.models.GHSOM [-h] [-l <labeling>] [-n <numberLabels>] [--numberWinners <numberWinners>] [--skipDWM] <properties> Options: -h Generate HTML output. -l <labeling> Labeling algorithm to use. -n <numberLabels> Number of labels per unit. Useless if no labeling algorithm is specified. --numberWinners <numberWinners> Number of winners to write. Default is 300. default: 300 --skipDWM Skip writing the data winner mapping file <properties> Name of property file.
GrowingSOM
Provides a Growing Grid (Bernd Fritzke, 1995), i.e. a Self-Organising Map that can dynamically grow by inserting whole rows or cells. When setting the growth-controlling parameters to disable growth, it can also be used to train a standard SOM.
Usage: java at.tuwien.ifs.somtoolbox.models.GrowingSOM [-h] [-l <labeling>] [-n <numberLabels>] [-w <weightVectorFile>] [-m <mapDescriptionFile>] [--skipDWM] [--numberWinners <numberWinners>] <properties> [--cpus <cpus>] Options: -h Generate HTML output. -l <labeling> Labeling algorithm to use. -n <numberLabels> Number of labels per unit. Useless if no labeling algorithm is specified. -w <weightVectorFile> Weight vector file used for initialization of the map. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. --skipDWM Skip writing the data winner mapping file --numberWinners <numberWinners> Number of winners to write. Default is 300. default: 300 <properties> Name of property file. --cpus <cpus> Numbers of CPUs to use. default: 1
MnemonicSOM
MenmonicSOM has a grid which is not fully occupied with units.
Usage: java at.tuwien.ifs.somtoolbox.models.MnemonicSOM [-h] [-l <labeling>] [-n <numberLabels>] --dim <dimension> -m <mapDescriptionFile> -u <unitDescriptionFile> [--numberWinners <numberWinners>] [--skipDWM] <properties> Options: -h Generate HTML output. -l <labeling> Labeling algorithm to use. -n <numberLabels> Number of labels per unit. Useless if no labeling algorithm is specified. --dim <dimension> The dimension of the input data. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. --numberWinners <numberWinners> Number of winners to write. Default is 300. default: 300 --skipDWM Skip writing the data winner mapping file <properties> Name of property file.
SOMTrainer
The SOMTrainer provides a graphical Interface to create a SOM based on different SOM Models.
Usage: java at.tuwien.ifs.somtoolbox.apps.trainer.SOMTrainer Options:
Viewer Applications
SOMViewer
An interactive viewer for exploring SOMs, using different visualisations
Usage: java at.tuwien.ifs.somtoolbox.apps.viewer.SOMViewer -u <unitDescriptionFile> -w <weightVectorFile> [--dw <dataWinnerMappingFile>] [-m <mapDescriptionFile>] [(-t|--tv) <templateVectorFile>] [-v <inputVectorFile>] [-c <classInformationFile>] [-r <regressionInformationFile>] [-i <dataInformationFile>] [-d <dataNamesFile>] [--linkage <linkageMapFile>] [--corrections <inputCorrections>] [--colours <classColours>] [-p <fileNamePrefix>] [-s <fileNameSuffix>] [--imagePrefix <imagePrefix>] [--imageSuffix <imageSuffix>] [--appdir <applicationDirectory>] [--datadir <viewerWorkingDirectory>] [--vis <initialVisualisation>] [--visParams <initialVisParams>] [--palette <initialPalette>] [(-2|--secondSOM) <secondSOMPrefix>] [-o|--documenMode] [--noplayer] [--pDdecode <probabilityDecode>] [--decodeDir <decodedOutputDir>] Options: -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. --dw <dataWinnerMappingFile> Unit description file describing the winners mapped onto a unit for a SOM/GHSOM. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. -t, --tv <templateVectorFile> Template vector file containing vector element labels. -v <inputVectorFile> Input file containing the input vectors to be mapped. -c <classInformationFile> Class information file containing the class for each data item. -r <regressionInformationFile> Regression information file containing the predicted values for each data item. -i <dataInformationFile> Data information file containing information such as location of each data item. -d <dataNamesFile> File containing the names of data items to be highlighted. --linkage <linkageMapFile> File containing data item linkage mapping. --corrections <inputCorrections> Name of the file containing input corrections. --colours <classColours> File listing the RGB values of the colours to be used for the class-based visualisations Pie-charts, Thematic class map. -p <fileNamePrefix> Prefix for the relative path to the files. This option is only to be used if no absolute paths are available. -s <fileNameSuffix> Suffix to the filenames aka file endings, e.g. ".mp3". This option is only to be used if no absolute paths are available. --imagePrefix <imagePrefix> Prefix to prepend to the labels to find the images --imageSuffix <imageSuffix> Suffix to append to the labels to find the images --appdir <applicationDirectory> Directory containing the SOMViewer application start file and the somviewer.prop file. --datadir <viewerWorkingDirectory> Directory containing the input data items needed for the visualisations. --vis <initialVisualisation> The name of the initial visualisation. --visParams <initialVisParams> Parameters for the initial visualisation. Currently only implemented for SmoothedDataHistograms visualisations. --palette <initialPalette> The name of the initial palette to be used. -2, --secondSOM <secondSOMPrefix> Prefix for the set of files representing second SOM -o, --documenMode Activates the Document mode, which hides the PlaySOM toolbar for exporting playlists and instead activates a document preview. For this to work properly, you have to specify the document path prefix denoting the base path for the files, e.g. 'file:///c:/somemap/files/', and the suffix, like .html. --noplayer Don't use the internal player PlaySOMPlayer, create the classic PlaySOMPanel --pDdecode <probabilityDecode> When using multi-channel audio playback Probability to decode a mp3 file to wav before playing. --decodeDir <decodedOutputDir> When using multi-channel audio playback Decoded mp3 files will be saved to and loaded from this directory.
UnitFileViewer
Plots a unit file, can be used especially to plot a 3D-SOM
Usage: java at.tuwien.ifs.somtoolbox.apps.UnitFileViewer -u <unitDescriptionFile> [--showLabels <showLabels>] [--verbose] Options: -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. --showLabels <showLabels> How many labels per unit should be displayed default = 0 default: 0 --verbose Be more verbose...
Utils Applications
AttendeeMapper
Writes an HTML output of the map, with highlighting certain data items
Usage: java at.tuwien.ifs.somtoolbox.output.AttendeeMapper -d <dataNamesFile> [-l <labeling>] -u <unitDescriptionFile> <htmlFile> Options: -d <dataNamesFile> File containing the names of data items to be highlighted. -l <labeling> Labeling algorithm to use. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. <htmlFile> Name of HTML file to write.
DataMapper
Maps inputs to an already trained SOM. If a unit-file is given, the data items are added to the loaded map, without a unti file the mapping starts with an empty map.
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.DataMapper -w <weightVectorFile> [-m <mapDescriptionFile>] -v <inputVectorFile> [-u <unitDescriptionFile>] [-c <classInformationFile>] [--classlist classList1:classList2:...:classListN ] [-l <labeling>] [-n <numberLabels>] [--numberWinners <numberWinners>] [--skipDWM] [<output>] Options: -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. -v <inputVectorFile> Input file containing the input vectors to be mapped. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -c <classInformationFile> Class information file containing the class for each data item. --classlist classList1:classList2:...:classListN A List of class names -l <labeling> Labeling algorithm to use. -n <numberLabels> Number of labels per unit. Useless if no labeling algorithm is specified. --numberWinners <numberWinners> Number of winners to write. Default is 300. default: 300 --skipDWM Skip writing the data winner mapping file <output> Name of the output file.
DataSetViewer
Usage: java at.tuwien.ifs.somtoolbox.apps.DataSetViewer [-v <inputVectorFile>] Options: -v <inputVectorFile> Input file containing the input vectors to be mapped.
HTMLOutputter
Creates an HTML representation of the Map. The representation displays a hit histogram, and shows the names of the mapped inputs.
Usage: java at.tuwien.ifs.somtoolbox.output.HTMLOutputter [--metric <metric>] [--normalized] [-v <inputVectorFile>] [--dense] [-l <labeling>] [--ignoreZero] [-n <numberLabels>] [(-t|--tv) <templateVectorFile>] [-w <weightVectorFile>] [-u <unitDescriptionFile>] [-m <mapDescriptionFile>] <htmlFile> Options: --metric <metric> Name of the metric to be used for distance calculation in input space. default: at.tuwien.ifs.somtoolbox.layers.metrics.L2Metric --normalized Set, if vectors are normalized to unit length. At the moment this option is not crucial. -v <inputVectorFile> Input file containing the input vectors to be mapped. --dense Set if input data vectors are densely populated. -l <labeling> Labeling algorithm to use. --ignoreZero Ignore labels with zero mean value and que. -n <numberLabels> Number of labels per unit. Useless if no labeling algorithm is specified. -t, --tv <templateVectorFile> Template vector file containing vector element labels. -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. <htmlFile> Name of HTML file to write.
RhythmPatternsMatrixVisualisationSaver
Provides a batch mode to save rhythm patterns matrix visualisation of all inputs in a vector file to image files.
Usage: java at.tuwien.ifs.feature.evaluation.RhythmPatternsMatrixVisualisationSaver -v <inputVectorFile> [--width <width>] [--type <filetype>] [(-g|--unitGrid) <unitGrid>] [(-o|--output) <basename>] Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. --width <width> The width of a unit. default: 10 --type <filetype> default: png -g, --unitGrid <unitGrid> Whether to draw the grid of units. default: true -o, --output <basename>
SOMMerger
Combines the weight vectors of one or more SOM maps to an input vector file
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.SOMMerger [--mode <mode>] [--skipConversion] [--mapSize <mapSize>] [--inputDir <inputDir>] <output> [maps1 maps2 ... mapsN] Options: --mode <mode> The merging mode to use, i.e. Union, Intersection, or a number of vectors the term occurrs. If no mode is provided, all possible combinations will be generated. default: All --skipConversion Skip conversion of map files to input vector files, if you already did that before. --mapSize <mapSize> The size of the map to be used to write the properties files, e.g. 4x5. If not specified, a default map size will be computed, depending on the number of input vectors. --inputDir <inputDir> Path to the input directory. <output> Name of the output file. maps1 maps2 ... mapsN Prefix for the SOM maps to merge i.e. file name w/o .unit/.wgt extension
SimilarityRetrieval
Performs similarity retrieval in a given database (vector file)
Usage: java at.tuwien.ifs.feature.evaluation.SimilarityRetrieval -v <inputVectorFile> -l <inputLabel> [(-n|--numberNeighbours) <numberNeighbours>] [--metric <metric>] Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. -l <inputLabel> Name/label of the input vector -n, --numberNeighbours <numberNeighbours> Number of neighbours to find. --metric <metric> Name of the metric to be used for distance calculation in input space. default: at.tuwien.ifs.somtoolbox.layers.metrics.L2Metric
SimilarityRetrievalGUI
Provides a graphical interface for similarity retrieval. Allows to load multiple vector files
Usage: java at.tuwien.ifs.feature.evaluation.SimilarityRetrievalGUI Options:
SimilarityRetrievalWriter
Performs similarity retrieval in a given database (vector file), and writes the results in one file per data itme
Usage: java at.tuwien.ifs.feature.evaluation.SimilarityRetrievalWriter -v <inputVectorFile> [(-n|--numberNeighbours) <numberNeighbours>] [--metric <metric>] --outputDir <outputDirectory> [--startIndex <startIndex>] [--numberItems <numberItems>] Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. -n, --numberNeighbours <numberNeighbours> Number of neighbours to find. default: 1000 --metric <metric> Name of the metric to be used for distance calculation in input space. default: at.tuwien.ifs.somtoolbox.layers.metrics.L1Metric --outputDir <outputDirectory> Name of the output directory. --startIndex <startIndex> The start index. default: 0 --numberItems <numberItems> Number of items
SomFilePacker
Packer to create autonomous SomFiles
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.SomFilePacker <output> -u <unitDescriptionFile> -w <weightVectorFile> [-d <dataNamesFile>] [-c <classInformationFile>] [-r <regressionInformationFile>] [-m <mapDescriptionFile>] [-i <dataInformationFile>] [-p <fileNamePrefix>] [-s <fileNameSuffix>] [--dw <dataWinnerMappingFile>] [-v <inputVectorFile>] [(-t|--tv) <templateVectorFile>] [--linkage <linkageMapFile>] [--colours <classColours>] [--corrections <inputCorrections>] Options: <output> Name of the output file. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. -d <dataNamesFile> File containing the names of data items to be highlighted. -c <classInformationFile> Class information file containing the class for each data item. -r <regressionInformationFile> Regression information file containing the predicted values for each data item. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. -i <dataInformationFile> Data information file containing information such as location of each data item. -p <fileNamePrefix> Prefix for the relative path to the files. This option is only to be used if no absolute paths are available. -s <fileNameSuffix> Suffix to the filenames aka file endings, e.g. ".mp3". This option is only to be used if no absolute paths are available. --dw <dataWinnerMappingFile> Unit description file describing the winners mapped onto a unit for a SOM/GHSOM. -v <inputVectorFile> Input file containing the input vectors to be mapped. -t, --tv <templateVectorFile> Template vector file containing vector element labels. --linkage <linkageMapFile> File containing data item linkage mapping. --colours <classColours> File listing the RGB values of the colours to be used for the class-based visualisations Pie-charts, Thematic class map. --corrections <inputCorrections> Name of the file containing input corrections.
TemplateVectorComparator
Compares the contents of two template vectors
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.TemplateVectorComparator <templateVectorFile> <templateVectorFile2> Options: <templateVectorFile> First template vector file. <templateVectorFile2> Second template vector file.
TrajectoryOutputter
Generates a graphical representation of a trajectory of the given points over the map
Usage: java at.tuwien.ifs.somtoolbox.output.TrajectoryOutputter -d <dataNamesFile> -u <unitDescriptionFile> [-m <mapDescriptionFile>] [-l|--drawLines] <imageFile> Options: -d <dataNamesFile> File containing the names of data items to be highlighted. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. -l, --drawLines Draw trajectory lines on map. <imageFile> Name of image file to write. No Suffix needed. Writes as PNG and EPS.
VisualisationImageSaver
Provides a batch mode to save several/all visualisations of a map to image files.
Usage: java at.tuwien.ifs.somtoolbox.apps.VisualisationImageSaver -u <unitDescriptionFile> -w <weightVectorFile> [-v <inputVectorFile>] [(-t|--tv) <templateVectorFile>] [--dw <dataWinnerMappingFile>] [-c <classInformationFile>] [(-o|--output) <basename>] [--width <width>] [--height <height>] [--type <filetype>] [(-g|--unitGrid) <unitGrid>] [vis1 vis2 ... visN] Options: -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. -v <inputVectorFile> Input file containing the input vectors to be mapped. -t, --tv <templateVectorFile> Template vector file containing vector element labels. --dw <dataWinnerMappingFile> Unit description file describing the winners mapped onto a unit for a SOM/GHSOM. -c <classInformationFile> Class information file containing the class for each data item. -o, --output <basename> --width <width> The width of a unit. default: 10 --height <height> The height of a unit. --type <filetype> default: png -g, --unitGrid <unitGrid> Whether to draw the grid of units. default: true vis1 vis2 ... visN The visualisations to create.
Helper Applications
ClassSubsetGenerator
Generates a subset and class assignment based on the given class file
Usage: java at.tuwien.ifs.feature.ClassSubsetGenerator -c <classInformationFile> [(-z|--gzip) <gzip>] <input> <output> Options: -c <classInformationFile> Class information file containing the class for each data item. -z, --gzip <gzip> Whether or not to gzip the output. default: true <input> Name of input vector file to be read. <output> Name of the output file.
DataWinnerMappingWriter
Writes a data winner mapping file from a trained map
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.DataWinnerMappingWriter -v <inputVectorFile> -w <weightVectorFile> -u <unitDescriptionFile> [-m <mapDescriptionFile>] --numberWinners <numberWinners> <output> [--outputDir <outputDirectory>] Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. --numberWinners <numberWinners> Number of winners to write. Default is 300. default: 300 <output> Name of the output file. --outputDir <outputDirectory> Name of the output directory.
DatasetRandomiser
Randomises data sets by swapping the order of columns (features/attributes) and/or rows (vectors)
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.DatasetRandomiser -v <inputVectorFile> [(-t|--tv) <templateVectorFile>] [--variants <variants>] [--interleave <interleave>] [--startIndex <startIndex>] [--preserveFeatureOrder] [--preserveVectorOrder] [(-z|--gzip) <gzip>] <output> Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. -t, --tv <templateVectorFile> Template vector file containing vector element labels. --variants <variants> Number of variants to generate. default: 1 --interleave <interleave> Interleave between the indices. default: 1 --startIndex <startIndex> The start index. default: 1 --preserveFeatureOrder Wether or not preserve the order of features. --preserveVectorOrder Wether or not preserve the order of vectors. -z, --gzip <gzip> Whether or not to gzip the output. default: true <output> Name of the output file.
DistanceMatrixWriter
Writes a distance matrix for the given data, in ASCII or binary format
Usage: java at.tuwien.ifs.somtoolbox.data.distance.DistanceMatrixWriter -v <inputVectorFile> [-c <classInformationFile>] [--metric <metric>] [--metricParams <metricParams>] <output> [--outputFormat <outputFormat>] Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. -c <classInformationFile> Class information file containing the class for each data item. --metric <metric> Name of the metric to be used for distance calculation in input space. default: at.tuwien.ifs.somtoolbox.layers.metrics.L2Metric --metricParams <metricParams> Parameters for the metric. <output> Name of the output file. --outputFormat <outputFormat> Format of the output file, valid values are: SOMLib, plain, Binary, Orange default: SOMLib
InputDataFileFormatConverter
Converts between various file formats for input data. Currently supported formats are [randomAccess, simpleMatrix, SOMPak, Marsyas0.2ARFF, SOMLib, ESOM, ARFF] and [SOMLib, ARFF, randomAccess, ESOM, SOMPak, Orange, CSV, vowpal], respective
Usage: java at.tuwien.ifs.somtoolbox.data.InputDataFileFormatConverter [--inputFormat <inputFormat>] <input> [(-t|--tv) <templateVectorFile>] [-c <classInformationFile>] [--outputFormat <outputFormat>] [(-z|--gzip) <gzip>] <output> [--skipInstanceNames] [--skipInputsWithoutClass] [--tabSeparated] Options: --inputFormat <inputFormat> Format of the input file, valid values are: randomAccess, simpleMatrix, SOMPak, Marsyas0.2ARFF, SOMLib, ESOM, ARFF If not specified, the format will be determined from the file extension. <input> Name of input vector file to be read. -t, --tv <templateVectorFile> Template vector file containing vector element labels. -c <classInformationFile> Class information file containing the class for each data item. --outputFormat <outputFormat> Format of the output file, valid values are: SOMLib, ARFF, randomAccess, ESOM, SOMPak, Orange, CSV, vowpal If not specified, the format will be determined from the file extension. -z, --gzip <gzip> Whether or not to gzip the output. default: true <output> Name of the output file. --skipInstanceNames Skipping writing of instance names in ARFF file --skipInputsWithoutClass Skipping writing instances without assigned class to ARFF file --tabSeparated Write the class-file tab separated.
LabelSOM
Implements the LabelSOM labelling method
Usage: java at.tuwien.ifs.somtoolbox.output.labeling.LabelSOM -v <inputVectorFile> (-t|--tv) <templateVectorFile> -w <weightVectorFile> -u <unitDescriptionFile> [-n <numberLabels>] [--dense] [--ignoreZero] [-m <mapDescriptionFile>] Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. -t, --tv <templateVectorFile> Template vector file containing vector element labels. -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -n <numberLabels> Number of labels per unit. Default value is 5. default: 5 --dense Set if input data vectors are densely populated. --ignoreZero Ignore labels with zero mean value and que. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM.
LagusKeywordLabeler
Implements the LagusKeyword labelling method
Usage: java at.tuwien.ifs.somtoolbox.output.labeling.LagusKeywordLabeler -v <inputVectorFile> (-t|--tv) <templateVectorFile> -w <weightVectorFile> -u <unitDescriptionFile> [-n <numberLabels>] [--dense] [-m <mapDescriptionFile>] --inputDir <inputDir> Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. -t, --tv <templateVectorFile> Template vector file containing vector element labels. -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -n <numberLabels> Number of labels per unit. Default value is 5. default: 5 --dense Set if input data vectors are densely populated. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. --inputDir <inputDir> Path to the input directory.
MapFileFormatConverter
Converts between various file formats for trained SOMs.Currently supported formats are [SOMLib, SOMPak, ESOM]
Usage: java at.tuwien.ifs.somtoolbox.input.MapFileFormatConverter --inputFormat <inputFormat> <input> [-v <inputVectorFile>] [(-t|--tv) <templateVectorFile>] [-u <unitDescriptionFile>] [--outputDir <outputDirectory>] --outputFormat <outputFormat> [(-z|--gzip) <gzip>] <output> Options: --inputFormat <inputFormat> Format of the input file, valid values are: SOMLib, SOMPak, ESOM <input> Name of input vector file to be read. -v <inputVectorFile> Input file containing the input vectors to be mapped. -t, --tv <templateVectorFile> Template vector file containing vector element labels. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. --outputDir <outputDirectory> Name of the output directory. --outputFormat <outputFormat> Format of the output file, valid values are: SOMLib, SOMPak, ESOM -z, --gzip <gzip> Whether or not to gzip the output. default: true <output> Name of the output file.
MapRotator
Rotates a map by the given degrees, and writes a new unit- and weight-vector file
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.MapRotator -u <unitDescriptionFile> -w <weightVectorFile> [--dw <dataWinnerMappingFile>] <output> [(-r|--rotation) <rotation>] [(-f|--flip) <flip>] Options: -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. --dw <dataWinnerMappingFile> Unit description file describing the winners mapped onto a unit for a SOM/GHSOM. <output> Name of the output file. -r, --rotation <rotation> Rotation of the new map, values are: 90, 180, 270. -f, --flip <flip> Flip the map, values are horizontal or vertical.
MnemonicSOMGenerator
MnemonicSOMGenerator allows to create an arbitrary shaped SOM (unit file)
Usage: java at.tuwien.ifs.somtoolbox.util.mnemonic.MnemonicSOMGenerator <backgroundImage> [(-n|--nodes) <totalNodes>] [(-r|--rows) <rows>] [(-c|--columns) <cols>] Options: <backgroundImage> The Background Image png, jpg -n, --nodes <totalNodes> -r, --rows <rows> -c, --columns <cols>
QualityMeasureComputer
Wrapper for the individual Quality Measures
Usage: java at.tuwien.ifs.somtoolbox.apps.QualityMeasureComputer -w <weightVectorFile> -m <mapDescriptionFile> -u <unitDescriptionFile> -v <inputVectorFile> [(-t|--tv) <templateVectorFile>] [--dw <dataWinnerMappingFile>] --qualityClass <qualityMeasureClass> --qualityVariant <qualityMeasureVariant> [-k <k>] <output> [<properties>] Options: -w <weightVectorFile> Weight vector file describing a SOM/GHSOM. -m <mapDescriptionFile> Map description file describing a mapped SOM/GHSOM. -u <unitDescriptionFile> Unit description file describing a mapped SOM/GHSOM. -v <inputVectorFile> Input file containing the input vectors to be mapped. -t, --tv <templateVectorFile> Template vector file containing vector element labels. --dw <dataWinnerMappingFile> Unit description file describing the winners mapped onto a unit for a SOM/GHSOM. --qualityClass <qualityMeasureClass> Quality measure class. --qualityVariant <qualityMeasureVariant> Quality measure variant. -k <k> <output> Name of the output file. <properties> Name of property file.
RGBPaletteConverter
Converts palettes given as RGB values to the XML format used in the SOMToolbox.
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.RGBPaletteConverter -f <inputFile> [--name <name>] [--shortName <shortName>] [--description <description>] <output> Options: -f <inputFile> Name of input file to be read. --name <name> Name of the palette. --shortName <shortName> Short name of the palette. --description <description> Description of the palette. <output> Name of the output file.
SOMLibInputConcatenator
Merges two or more SOMLib Input files, i.e. vector and template files. Template vectors can be off different dimensionality, and may contain different features, but some features may also be overlapping. Different merge strategies are available: union of all features sets, intersection of features sets, and strategies in between, retaining a feature if it appears in at least x sets
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.SOMLibInputConcatenator [--mode <mode>] [--inputDir <inputDir>] <output> [inputs1 inputs2 ... inputsN] Options: --mode <mode> The merging mode to use, i.e. Union, Intersection, or a number of vectors the term occurrs. If no mode is provided, all possible combinations will be generated. default: All --inputDir <inputDir> Path to the input directory. <output> Name of the output file. inputs1 inputs2 ... inputsN Prefix for the input files to merge i.e. file name w/o .tv/.vec extension
SOMLibVectorNormalization
Handles the normalization of vector files in SOMLib format
Usage: java at.tuwien.ifs.somtoolbox.data.SOMLibVectorNormalization [-m <method>] <input> <output> Options: -m <method> Normalization method. UNIT_LEN normalises the vectors in the input file to unit length. MIN_MAX normalises each attributes between 0 and 1. STANDARD_SCORE normalises each attribute to a mean of 0, and a max value of the standard deviation. default: UNIT_LEN <input> Name of input vector file to be read. <output> Name of new vector file to be created.
SOMLibZeroVectorRemover
Removes zero vectors (i.e. with 0 in all their components) from files in SOMLib format
Usage: java at.tuwien.ifs.somtoolbox.data.SOMLibZeroVectorRemover <input> <output> Options: <input> Name of input vector file to be read. <output> Name of new vector file to be created.
StringReplacer
Replaces strings by a replacement in the given file
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.StringReplacer --replace <replace> [--replacement <replacement>] [-f <inputFile>] [--inputDir <inputDir>] Options: --replace <replace> The string to be replace. --replacement <replacement> The string to replace with. -f <inputFile> Name of input file to be read. --inputDir <inputDir> Path to the input directory.
VectorFile2DatabaseImporter
Imports input and template vector files to a database
Usage: java at.tuwien.ifs.somtoolbox.database.VectorFile2DatabaseImporter -v <inputVectorFile> (-t|--tv) <templateVectorFile> --dbName <databaseName> --tablePrefix <databaseTableNamePrefix> [--server <databaseServerAddress>] [--user <databaseUser>] [--password <databasePassword>] Options: -v <inputVectorFile> Input file containing the input vectors to be mapped. -t, --tv <templateVectorFile> Template vector file containing vector element labels. --dbName <databaseName> Name of the database. --tablePrefix <databaseTableNamePrefix> Prefix for the tables in the database. --server <databaseServerAddress> Servername or IP of the database server. Defaults to 'localhost'. default: localhost --user <databaseUser> Username for the database acccess. Defaults to 'root'. default: root --password <databasePassword> Password for the database acccess. Defaults to an empty password. default:
VectorFileAppender
Append multilple VectorFiles containing the same type of features into one Vector file
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.VectorFileAppender <output> input1 input2 ... inputN Options: <output> Name of new vector file to be created. input1 input2 ... inputN The input files
VectorFileConcatenator
Merge two or more VectorFiles containing different Features of the same Data into one Vector file
Usage: java at.tuwien.ifs.somtoolbox.apps.helper.VectorFileConcatenator <output> input1 input2 ... inputN [(-q|--weights) weights1:weights2:...:weightsN ] [--outputFormat <outputFormat>] [--tv] Options: <output> Name of new vector file to be created. input1 input2 ... inputN The input files -q, --weights weights1:weights2:...:weightsN Apply different weights when normalising the vectors. No normalisation if skipped, missing values default to 1 --outputFormat <outputFormat> Format of the output file, valid values are: SOMLib, ARFF, randomAccess, ESOM, SOMPak, Orange, CSV, vowpal If not specified, the format will be determined from the file extension. --tv Create and write an apropriate TemplateVector file: <outfile>.tv
VectorFileRewriter
Replaces labels in SOMLibVectorFiles by means of a mapping file; labels in the mapping file must be separated by a tab