ELKI command line parameter overview by option:

-abod.kernelfunction <class|object>

Kernel function to use.

Class Restriction: implements distance.similarityfunction.SimilarityFunction

Default: kernel.PolynomialKernelFunction

Known implementations:

Parameter for:

-abod.l <int>

Number of top outliers to compute.

Parameter for:

-achlioptas.sparsity <double>

Frequency of zeros in the projection matrix.

Default: 3.0

Parameter for:

-adapter.similarityfunction <class|object>

Similarity function to derive the distance between database objects from.

Class Restriction: implements distance.similarityfunction.NormalizedSimilarityFunction

Known implementations:

Parameter for:

-algorithm <object_1|class_1,...,object_n|class_n>

Algorithm to run.

Class Restriction: implements algorithm.Algorithm

Known implementations:

Parameter for:

-algorithm.distancefunction <class|object>

Distance function to determine the distance between database objects.

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-ap.convergence <int>

Number of stable iterations for convergence.

Default: 15

Parameter for:

-ap.distance <class|object>

Distance function to use.

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.SquaredEuclideanDistanceFunction

Known implementations:

Parameter for:

-ap.initialization <class|object>

Similarity matrix initialization..

Class Restriction: implements algorithm.clustering.affinitypropagation.AffinityPropagationInitialization

Default: DistanceBasedInitializationWithMedian

Known implementations:

Parameter for:

-ap.lambda <double>

Dampening factor lambda. Usually 0.5 to 1.

Default: 0.5

Parameter for:

-ap.maxiter <int>

Maximum number of iterations.

Default: 1000

Parameter for:

-ap.quantile <double>

Quantile to use for diagonal entries.

Default: 0.5

Parameter for:

-ap.similarity <class|object>

Similarity function to use.

Class Restriction: implements distance.similarityfunction.SimilarityFunction

Default: kernel.LinearKernelFunction

Known implementations:

Parameter for:

-arff.classlabel <pattern>

Pattern to recognize class label attributes.

Default: (Class|Class-?Label)

Parameter for:

-arff.externalid <pattern>

Pattern to recognize external ID attributes.

Default: (External-?ID)

Parameter for:

-associationrules.algorithm <class|object>

Algorithm to be used for frequent itemset mining.

Class Restriction: extends algorithm.itemsetmining.AbstractFrequentItemsetAlgorithm

Default: FPGrowth

Known implementations:

Parameter for:

-associationrules.interestingness <class|object>

Interestingness measure to be used

Class Restriction: implements algorithm.itemsetmining.associationrules.interest.InterestingnessMeasure

Default: Confidence

Known implementations:

Parameter for:

-associationrules.maxmeasure <double>

Maximum threshold for specified interstingness measure

Parameter for:

-associationrules.minmeasure <double>

Minimum threshold for specified interstingness measure

Parameter for:

-autopca.filter <class|object>

Filter for selecting eigenvectors during autotuning PCA.

Class Restriction: implements math.linearalgebra.pca.filter.EigenPairFilter

Default: SignificantEigenPairFilter

Known implementations:

Parameter for:

-avep.includeself <|true|false>

Include the query object in the evaluation.

Default: false

Parameter for:

-avep.k <int>

K to compute the average precision at.

Parameter for:

-avep.sampling <double>

Relative amount of object to sample.

Parameter for:

-avep.sampling-seed <long>

Random seed for deterministic sampling.

Default: use global random seed

Parameter for:

-axisref.scale <double>

Scale the data space extension by the given factor.

Default: 1.0

Parameter for:

-ay.k <int>

Subspace dimensionality to search for.

Parameter for:

-ay.m <int>

Population size for evolutionary algorithm.

Parameter for:

-ay.phi <int>

The number of equi-depth grid ranges to use in each dimension.

Parameter for:

-ay.seed <long>

The random number generator seed.

Default: use global random seed

Parameter for:

-bisecting.kmeansvariant <class|object>

KMeans variant

Class Restriction: implements algorithm.clustering.kmeans.KMeans

Default: BestOfMultipleKMeans

Known implementations:

Parameter for:

-bubble.fill <|true|false>

Half-transparent filling of bubbles.

Default: false

Parameter for:

-bubble.scaling <class|object>

Additional scaling function for bubbles.

Class Restriction: implements utilities.scaling.ScalingFunction

Default: outlier.OutlierLinearScaling

Known implementations:

Parameter for:

-bundle.input <file>

Bundle file to load the data from.

Parameter for:

-bylabelclustering.multiple <|true|false>

Flag to indicate that only subspaces with large coverage (i.e. the fraction of the database that is covered by the dense units) are selected, the rest will be pruned.

Default: false

Parameter for:

-bylabelclustering.noise <pattern>

Pattern to recognize noise classes by their label.

Parameter for:

-bymodel.noise <pattern>

Pattern to recognize noise models by their label.

Parameter for:

-bymodel.randomseed <long>

The random generator seed.

Default: use global random seed

Parameter for:

-bymodel.reassign <pattern>

Pattern to specify clusters to reassign.

Parameter for:

-bymodel.sizescale <double>

Factor for scaling the specified cluster sizes.

Default: 1.0

Parameter for:

-bymodel.spec <file>

The generator specification file.

Parameter for:

-c-index.distance <class|object>

Distance function to use for computing the c-index.

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-c-index.noisehandling <MERGE_NOISE | TREAT_NOISE_AS_SINGLETONS | IGNORE_NOISE>

Control how noise should be treated.

Default: TREAT_NOISE_AS_SINGLETONS

Parameter for:

-canopy.t1 <double>

Inclusion threshold for canopy clustering. t1 >= t2!

Parameter for:

-canopy.t2 <double>

Removal threshold for canopy clustering. t1 >= t2!

Parameter for:

-cash.adjust <|true|false>

Flag to indicate that an adjustment of the applied heuristic for choosing an interval is performed after an interval is selected.

Default: false

Parameter for:

-cash.jitter <double>

The maximum jitter for distance values.

Parameter for:

-cash.maxlevel <int>

The maximum level for splitting the hypercube.

Parameter for:

-cash.mindim <int>

The minimum dimensionality of the subspaces to be found.

Default: 1

Parameter for:

-cash.minpts <int>

Threshold for minimum number of points in a cluster.

Parameter for:

-cblof.algorithm <class|object>

Clustering algorithm to use for detecting outliers.

Class Restriction: implements algorithm.clustering.ClusteringAlgorithm

Default: kmeans.KMeansSort

Known implementations:

Parameter for:

-cblof.alpha <double>

The ratio of the data that should be included in the large clusters

Parameter for:

-cblof.beta <double>

The ratio of the data that should be included in the large clusters

Parameter for:

-cftree.absorption <class|object>

Absorption criterion to use.

Class Restriction: implements algorithm.clustering.hierarchical.birch.BIRCHAbsorptionCriterion

Default: DiameterCriterion

Known implementations:

Parameter for:

-cftree.branching <int>

Maximum branching factor of the CF-Tree

Default: 64

Parameter for:

-cftree.distance <class|object>

Distance function to use for node assignment.

Class Restriction: implements algorithm.clustering.hierarchical.birch.BIRCHDistance

Default: VarianceIncreaseDistance

Known implementations:

Parameter for:

-cftree.maxleaves <double>

Maximum number of leaves (if less than 1, the values is assumed to be relative)

Default: 0.05

Parameter for:

-cftree.threshold <double>

Threshold for adding points to existing nodes in the CF-Tree.

Parameter for:

-changepointdetection.bootstrap.confidence <double>

Confidence level to use with bootstrap sampling.

Default: 0.9975

Parameter for:

-changepointdetection.bootstrap.samples <int>

Number of samples to draw for bootstrapping the confidence estimate.

Default: 1000

Parameter for:

-changepointdetection.seed <long>

Random generator seed for bootstrap sampling.

Default: use global random seed

Parameter for:

-chengandchurch.alpha <double>

Parameter for multiple node deletion to accelerate the algorithm.

Default: 1.0

Parameter for:

-chengandchurch.delta <double>

Threshold value to determine the maximal acceptable score (mean squared residue) of a bicluster.

Parameter for:

-chengandchurch.n <int>

The number of biclusters to be found.

Default: 1

Parameter for:

-chengandchurch.replacement <class|object>

Distribution of replacement values when masking found clusters.

Class Restriction: implements math.statistics.distribution.Distribution

Default: UniformDistribution

Known implementations:

Parameter for:

-clara.independent <|true|false>

Draw independent samples (default is to keep the previous best medoids in the sample).

Default: false

Parameter for:

-clara.numlocal <int>

Number of samples (restarts) to run.

Default: 2

Parameter for:

-clara.numneighbor <double>

Number of tries to find a neighbor.

Default: 0.0125

Parameter for:

-clara.random <long>

Random generator seed.

Default: use global random seed

Parameter for:

-clara.samples <int>

Number of samples (iterations) to run.

Default: 5

Parameter for:

-clara.samplesize <double>

The size of the sample.

Default: 40.0

Parameter for:

-clarans.random <long>

Random generator seed.

Default: use global random seed

Parameter for:

-class.negative <string>

Class label to use for negative instances.

Default: negative

Parameter for:

-class.pattern <pattern>

Regular expression to identify positive objects.

Parameter for:

-class.positive <string>

Class label to use for positive instances.

Default: positive

Parameter for:

-clipscale.max <double>

Maximum value to allow.

Parameter for:

-clipscale.min <double>

Minimum value to allow.

Parameter for:

-clique.prune <|true|false>

Flag to indicate that only subspaces with large coverage (i.e. the fraction of the database that is covered by the dense units) are selected, the rest will be pruned.

Default: false

Parameter for:

-clique.tau <double>

The density threshold for the selectivity of a unit, where the selectivity is the fraction of total feature vectors contained in this unit.

Parameter for:

-clique.xsi <int>

The number of intervals (units) in each dimension.

Parameter for:

-clustering.label <string>

Parameter to override the clustering label, mostly to give a more descriptive label.

Parameter for:

-clustering.output <file>

Output file name. When not given, the result will be written to stdout.

Parameter for:

-clustering.output.append <|true|false>

Always append to the output file.

Default: false

Parameter for:

-cof.k <int>

The number of neighbors (not including the query object) to use for computing the COF score.

Parameter for:

-comphist.bins <int>

number of bins

Default: 50

Parameter for:

-comphist.positive <pattern>

Class label for the 'positive' class.

Parameter for:

-comphist.positive <pattern>

Class label for the 'positive' class.

Parameter for:

-comphist.scaling <class|object>

Class to use as scaling function.

Class Restriction: implements utilities.scaling.ScalingFunction

Default: IdentityScaling

Known implementations:

Parameter for:

-comphist.scaling <class|object>

Class to use as scaling function.

Class Restriction: implements utilities.scaling.ScalingFunction

Default: IdentityScaling

Known implementations:

Parameter for:

-concordant-pairs.noisehandling <MERGE_NOISE | TREAT_NOISE_AS_SINGLETONS | IGNORE_NOISE>

Control how noise should be treated.

Default: TREAT_NOISE_AS_SINGLETONS

Parameter for:

-concordant.distance <class|object>

Distance function to use for measuring concordant and discordant pairs.

Class Restriction: implements distance.distancefunction.PrimitiveDistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-cop.dist <CHISQUARED | GAMMA>

The assumed distribution of squared distances. ChiSquared is faster, Gamma expected to be more accurate but could also overfit.

Default: GAMMA

Parameter for:

-cop.expect <double>

Expected share of outliers. Only affect score normalization.

Default: 0.001

Parameter for:

-cop.k <int>

The number of nearest neighbors of an object to be considered for computing its COP_SCORE.

Parameter for:

-cop.k <int>

The number of nearest neighbors of an object to be considered for computing its COP_SCORE.

Parameter for:

-cop.models <|true|false>

Include COP models (error vectors) in output. This needs more memory.

Default: false

Parameter for:

-cop.pcarunner <class|object>

The class to compute (filtered) PCA.

Class Restriction: extends math.linearalgebra.pca.PCARunner

Default: PCARunner

Known implementations:

Parameter for:

-cop.pcarunner <class|object>

The class to compute (filtered) PCA.

Class Restriction: extends math.linearalgebra.pca.PCARunner

Default: PCARunner

Known implementations:

Parameter for:

-copac.knn <int>

Number of neighbors to use for PCA.

Parameter for:

-copscaling.phi <double>

Phi parameter, expected rate of outliers. Set to 0 to use raw CDF values.

Parameter for:

-covertree.distancefunction <class|object>

Distance function to determine the distance between objects.

Class Restriction: implements distance.distancefunction.DistanceFunction

Known implementations:

Parameter for:

-covertree.expansionrate <double>

Expansion rate of the tree (Default: 1.3).

Default: 1.3

Parameter for:

-covertree.truncate <int>

Truncate tree when branches have less than this number of instances.

Default: 10

Parameter for:

-davies-bouldin.distance <class|object>

Distance function to use for computing the davies-bouldin index.

Class Restriction: implements distance.distancefunction.NumberVectorDistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-davies-bouldin.noisehandling <MERGE_NOISE | TREAT_NOISE_AS_SINGLETONS | IGNORE_NOISE>

Control how noise should be treated.

Default: TREAT_NOISE_AS_SINGLETONS

Parameter for:

-db <class|object>

Database class.

Class Restriction: implements database.Database

Default: StaticArrayDatabase

Known implementations:

Parameter for:

-db.index <object_1|class_1,...,object_n|class_n>

Database indexes to add.

Class Restriction: implements index.IndexFactory

Known implementations:

Parameter for:

-dbc <class|object>

Database connection class.

Class Restriction: implements datasource.DatabaseConnection

Default: FileBasedDatabaseConnection

Known implementations:

Parameter for:

-dbc.classLabelClass <class|object>

Class label class to use.

Class Restriction: extends data.ClassLabel

Default: SimpleClassLabel

Known implementations:

Parameter for:

-dbc.classLabelIndex <int>

The index of the label to be used as class label. The first label is 0, negative indexes are relative to the end.

Parameter for:

-dbc.dim <int>

Dimensionality of the vectors to generate.

Parameter for:

-dbc.externalIdIndex <int>

The index of the label to be used as external Id. The first label is 0; negative indexes are relative to the end.

Parameter for:

-dbc.filter <object_1|class_1,...,object_n|class_n>

The filters to apply to the input data.

Class Restriction: implements datasource.filter.ObjectFilter

Known implementations:

Parameter for:

-dbc.genseed <long>

Seed for randomly generating vectors

Default: use global random seed

Parameter for:

-dbc.in <file_1,...,file_n>

The name of the input file to be parsed.

Parameter for:

-dbc.inputstream <class|object>

Input stream to read. Defaults to standard input.

Class Restriction: extends java.io.InputStream

Parameter for:

-dbc.parser <class|object>

Parser to provide the database.

Class Restriction: implements datasource.parser.Parser

Default: NumberVectorLabelParser

Known implementations:

Parameter for:

-dbc.size <int>

Database size to generate.

Parameter for:

-dbc.startid <int>

Object ID to start counting with

Default: 0

Parameter for:

-dbcv.distance <class|object>

Distance function to use for computing the dbcv.

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-dbod.d <double>

size of the D-neighborhood

Parameter for:

-dbod.p <double>

minimum fraction of objects that must be outside the D-neighborhood of an outlier

Parameter for:

-dbscan.epsilon <double>

The maximum radius of the neighborhood to be considered.

Parameter for:

-dbscan.minpts <int>

Threshold for minimum number of points in the epsilon-neighborhood of a point. The suggested value is '2 * dim - 1'.

Parameter for:

-deliclu.minpts <int>

Threshold for minimum number of points within a cluster.

Parameter for:

-dendrogram.layout <HALF_POS | HALF_WIDTH>

Positioning logic for dendrograms.

Default: HALF_POS

Parameter for:

-dendrogram.style <RECTANGULAR | TRIANGULAR_MAX | TRIANGULAR>

Drawing style for dendrograms.

Default: RECTANGULAR

Parameter for:

-derivator.accuracy <int>

Threshold for output accuracy fraction digits.

Default: 4

Parameter for:

-derivator.randomSample <|true|false>

Flag to use random sample (use knn query around centroid, if flag is not set).

Default: false

Parameter for:

-derivator.sampleSize <int>

Threshold for the size of the random sample to use. Default value is size of the complete dataset.

Parameter for:

-dim <int>

an integer between 1 and the dimensionality of the feature space 1 specifying the dimension to be considered for distance computation.

Parameter for:

-dish.epsilon <double_1,...,double_n>

A comma separated list of positive doubles specifying the maximum radius of the neighborhood to be considered in each dimension for determination of the preference vector (default is 0.001 in each dimension). If only one value is specified, this value will be used for each dimension.

Default: 0.001

Parameter for:

-dish.epsilon <double>

The maximum radius of the neighborhood to be considered in each dimension for determination of the preference vector.

Default: 0.001

Parameter for:

-dish.minpts <int>

Positive threshold for minumum numbers of points in the epsilon-neighborhood of a point. The value of the preference vector in dimension d_i is set to 1 if the epsilon neighborhood contains more than dish.minpts points and the following condition holds: for all dimensions d_j: |neighbors(d_i) intersection neighbors(d_j)| >= dish.minpts.

Parameter for:

-dish.mu <int>

The minimum number of points as a smoothing factor to avoid the single-link-effekt.

Default: 1

Parameter for:

-dish.strategy <APRIORI | MAX_INTERSECTION>

The strategy for determination of the preference vector, available strategies are: [APRIORI| MAX_INTERSECTION](default is MAX_INTERSECTION)

Default: MAX_INTERSECTION

Parameter for:

-distance.default <double>

Default distance to use for undefined values.

Default: Infinity

Parameter for:

-distance.dims <int_1,...,int_n>

a comma separated array of integer values, where 0 <= d_i < the dimensionality of the feature space specifying the dimensions to be considered for distance computation. If this parameter is not set, no dimensions will be considered, i.e. the distance between two objects is always 0.

Parameter for:

-distance.latitudedim <int>

The dimension containing the latitude.

Parameter for:

-distance.longitudedim <int>

The dimension containing the longitude.

Parameter for:

-distance.matrix <file>

The name of the file containing the distance matrix.

Parameter for:

-distance.matrix <file>

The name of the file containing the distance matrix.

Parameter for:

-distance.parser <class|object>

Parser used to load the distance matrix.

Class Restriction: implements distance.distancefunction.external.DistanceParser

Default: AsciiDistanceParser

Known implementations:

Parameter for:

-distance.weights <double_1,...,double_n>

Weights to use for the distance function.

Parameter for:

-distancefunction.index <class|object>

Distance index to use.

Class Restriction: implements index.preprocessed.snn.SharedNearestNeighborIndex

Default: SharedNearestNeighborPreprocessor

Known implementations:

Parameter for:

-distribution.beta.alpha <double>

Beta distribution alpha parameter

Parameter for:

-distribution.beta.beta <double>

Beta distribution beta parameter

Parameter for:

-distribution.cauchy.shape <double>

Cauchy distribution gamma/shape parameter.

Parameter for:

-distribution.chi.dof <double>

Chi distribution degrees of freedom parameter.

Parameter for:

-distribution.constant <double>

Constant value.

Parameter for:

-distribution.expgamma.shift <double>

Shift offset parameter.

Default: 0.0

Parameter for:

-distribution.exponential.rate <double>

Exponential distribution rate (lambda) parameter (inverse of scale).

Parameter for:

-distribution.gamma.k <double>

Gamma distribution k = alpha parameter.

Parameter for:

-distribution.gamma.theta <double>

Gamma distribution theta = 1/beta parameter.

Parameter for:

-distribution.kappa.shape1 <double>

First shape parameter of kappa distribution.

Parameter for:

-distribution.kappa.shape2 <double>

Second shape parameter of kappa distribution.

Parameter for:

-distribution.laplace.rate <double>

Laplace distribution rate (lambda) parameter (inverse of scale).

Parameter for:

-distribution.location <double>

Distribution location parameter

Parameter for:

-distribution.loggamma.shift <double>

Shift offset parameter.

Parameter for:

-distribution.lognormal.logmean <double>

Mean of the distribution before logscaling.

Parameter for:

-distribution.lognormal.logstddev <double>

Standard deviation of the distribution before logscaling.

Parameter for:

-distribution.lognormal.shift <double>

Shifting offset, so the distribution does not begin at 0.

Default: 0.0

Parameter for:

-distribution.max <double>

Maximum value of distribution.

Parameter for:

-distribution.min <double>

Minimum value of distribution.

Parameter for:

-distribution.poisson.n <int>

Number of trials.

Parameter for:

-distribution.poisson.probability <double>

Success probability.

Parameter for:

-distribution.random <long>

Random generation data source.

Default: use global random seed

Parameter for:

-distribution.scale <double>

Distribution scale parameter

Parameter for:

-distribution.shape <double>

Distribution shape parameter

Parameter for:

-distribution.skewgnormal.skew <double>

Skew of the distribution.

Parameter for:

-distribution.studentst.nu <int>

Degrees of freedom.

Parameter for:

-distsample.nozeros <|true|false>

Ignore zero distances, beneficial for data sets with many duplicates.

Default: false

Parameter for:

-distsample.quantile <double>

Quantile to compute.

Default: 0.1

Parameter for:

-distsample.sample <double>

Number of distances to compute, either relative (values less than 1), or absolute.

Parameter for:

-distsample.seed <long>

Random generator seed.

Default: use global random seed

Parameter for:

-diststat.bins <int>

Number of bins to use in the histogram. By default, it is only guaranteed to be within 1*n and 2*n of the given number.

Default: 20

Parameter for:

-diststat.exact <|true|false>

In a first pass, compute the exact minimum and maximum, at the cost of O(2*n*n) instead of O(n*n). The number of resulting bins is guaranteed to be as requested.

Default: false

Parameter for:

-diststat.sampling <|true|false>

Enable sampling of O(n) size to determine the minimum and maximum distances approximately. The resulting number of bins can be larger than the given n.

Default: false

Parameter for:

-doc.alpha <double>

Minimum relative density for a set of points to be considered a cluster (|C|>=doc.alpha*|S|).

Default: 0.2

Parameter for:

-doc.beta <double>

Preference of cluster size versus number of relevant dimensions (higher value means higher priority on larger clusters).

Default: 0.8

Parameter for:

-doc.random-seed <long>

Random seed, for reproducible experiments.

Default: use global random seed

Parameter for:

-doc.w <double>

Maximum extent of scattering of points along a single attribute for the attribute to be considered relevant.

Default: 0.05

Parameter for:

-dwof.delta <double>

Radius increase factor.

Default: 1.1

Parameter for:

-dwof.k <int>

Number of neighbors to get for DWOF score outlier detection.

Parameter for:

-edit.bandsize <double>

The band size for time series alignment. By default, no constraint is used. If the value is larger than 0, it will be considered absolute, otherwise relative to the longer sequence. Note that 0 does not make sense: use Euclidean distance then instead.

Parameter for:

-edr.delta <double>

the delta parameter (similarity threshold) for EDR (positive number)

Default: 1.0

Parameter for:

-em.centers <class|object>

Method to choose the initial cluster centers.

Class Restriction: implements algorithm.clustering.kmeans.initialization.KMeansInitialization

Default: RandomlyChosenInitialMeans

Known implementations:

Parameter for:

-em.delta <double>

The termination criterion for maximization of E(M): E(M) - E(M') < em.delta

Default: 1.0E-7

Parameter for:

-em.k <int>

The number of clusters to find.

Parameter for:

-em.map.prior <double>

Regularization factor for MAP estimation.

Parameter for:

-em.model <class|object>

Model factory.

Class Restriction: implements algorithm.clustering.em.EMClusterModelFactory

Default: MultivariateGaussianModelFactory

Known implementations:

Parameter for:

-enableDebug <string>

Parameter to enable debugging for particular packages.

Parameter for:

-ensemble.median.quantile <double>

Quantile to use in median voting.

Default: 0.5

Parameter for:

-ensemble.voting <class|object>

Voting strategy to use in the ensemble.

Class Restriction: implements utilities.ensemble.EnsembleVoting

Known implementations:

Parameter for:

-eric.k <int>

Number of neighbors to use for PCA.

Parameter for:

-ericdf.delta <double>

Threshold for approximate linear dependency: the strong eigenvectors of q are approximately linear dependent from the strong eigenvectors p if the following condition holds for all stroneg eigenvectors q_i of q (lambda_q < lambda_p): q_i' * M^check_p * q_i <= delta^2.

Default: 0.1

Parameter for:

-ericdf.tau <double>

Threshold for the maximum distance between two approximately linear dependent subspaces of two objects p and q (lambda_q < lambda_p) before considering them as parallel.

Default: 0.1

Parameter for:

-erp.g <double>

The g parameter of ERP - comparison value to use in gaps.

Default: 0.0

Parameter for:

-evaluator <object_1|class_1,...,object_n|class_n>

Class to evaluate the results with.

Class Restriction: implements evaluation.Evaluator

Default: AutomaticEvaluation

Known implementations:

Parameter for:

-extendedneighbors.neighborhood <class|object>

The inner neighborhood predicate to use.

Class Restriction: implements algorithm.outlier.spatial.neighborhood.NeighborSetPredicate

Known implementations:

Parameter for:

-extendedneighbors.neighborhood <class|object>

The inner neighborhood predicate to use.

Class Restriction: implements algorithm.outlier.spatial.neighborhood.NeighborSetPredicate

Known implementations:

Parameter for:

-extendedneighbors.steps <int>

The number of steps allowed in the neighborhood graph.

Parameter for:

-extendedneighbors.steps <int>

The number of steps allowed in the neighborhood graph.

Parameter for:

-external.knnfile <file>

Filename with the precomputed k nearest neighbors.

Parameter for:

-externalcluster.file <file>

The file name containing the (external) cluster vector.

Parameter for:

-externalneighbors.file <file>

The file listing the neighbors.

Parameter for:

-externaloutlier.file <file>

The file name containing the (external) outlier scores.

Parameter for:

-externaloutlier.idpattern <pattern>

The pattern to match object ID prefix

Default: ^ID=

Parameter for:

-externaloutlier.inverted <|true|false>

Flag to signal an inverted outlier score.

Default: false

Parameter for:

-externaloutlier.scaling <class|object>

Class to use as scaling function.

Class Restriction: implements utilities.scaling.ScalingFunction

Default: IdentityScaling

Known implementations:

Parameter for:

-externaloutlier.scorepattern <pattern>

The pattern to match object score prefix

Parameter for:

-extract.k <int>

The number of clusters to extract.

Parameter for:

-extract.minclsize <int>

The minimum cluster size.

Parameter for:

-farthest.keepfirst <|true|false>

Keep the first object chosen (which is chosen randomly) for the farthest points heuristic.

Default: false

Parameter for:

-fastabod.k <int>

Number of nearest neighbors to use for ABOD.

Parameter for:

-fastdoc.d0 <int>

Parameter for FastDOC, setting the number of relevant attributes which, when found for a cluster, are deemed enough to stop iterating.

Default: 5

Parameter for:

-fastoptics.randomproj.seed <long>

Random seed for generating projections.

Default: use global random seed

Parameter for:

-fbagging.breadth <|true|false>

Use the breadth first combinations instead of the cumulative sum approach

Default: false

Parameter for:

-fbagging.num <int>

The number of instances to use in the ensemble.

Parameter for:

-fbagging.seed <long>

Specify a particular random seed.

Default: use global random seed

Parameter for:

-fdbscan.samplesize <int>

The number of samples to draw from each uncertain object to determine the epsilon-neighborhood.

Parameter for:

-fdbscan.seed <long>

Random generator used to draw samples.

Default: use global random seed

Parameter for:

-fdbscan.threshold <double>

The amount of samples that have to be epsilon-close for two objects to be neighbors.

Default: 0.5

Parameter for:

-filter.dim <int>

Dimensionality of vectors to retain.

Parameter for:

-first.n <int>

Number of objects to keep.

Parameter for:

-gammascale.normalize <|true|false>

Regularize scores before using Gamma scaling.

Default: false

Parameter for:

-gaussod.invert <|true|false>

Invert the value range to [0:1], with 1 being outliers instead of 0.

Default: false

Parameter for:

-gdbscan.core <class|object>

Core point predicate for Generalized DBSCAN

Class Restriction: implements algorithm.clustering.gdbscan.CorePredicate

Default: MinPtsCorePredicate

Known implementations:

Parameter for:

-gdbscan.core <class|object>

Core point predicate for Generalized DBSCAN

Class Restriction: implements algorithm.clustering.gdbscan.CorePredicate

Default: MinPtsCorePredicate

Known implementations:

Parameter for:

-gdbscan.core-model <|true|false>

Use a model that keeps track of core points. Needs more memory.

Default: false

Parameter for:

-gdbscan.core-model <|true|false>

Use a model that keeps track of core points. Needs more memory.

Default: false

Parameter for:

-gdbscan.minsim <double>

Minimum similarity of points to cluster.

Parameter for:

-gdbscan.neighborhood <class|object>

Neighborhood predicate for Generalized DBSCAN

Class Restriction: implements algorithm.clustering.gdbscan.NeighborPredicate

Default: EpsilonNeighborPredicate

Known implementations:

Parameter for:

-gdbscan.neighborhood <class|object>

Neighborhood predicate for Generalized DBSCAN

Class Restriction: implements algorithm.clustering.gdbscan.NeighborPredicate

Default: EpsilonNeighborPredicate

Known implementations:

Parameter for:

-gdbscan.similarity <class|object>

Similarity function to use.

Class Restriction: implements distance.similarityfunction.SimilarityFunction

Known implementations:

Parameter for:

-generate.n <int>

The number of reference points to be generated.

Parameter for:

-generate.random <long>

Random generator seed.

Default: use global random seed

Parameter for:

-generate.scale <double>

Scale the grid by the given factor. This can be used to obtain reference points outside the used data space.

Default: 1.0

Parameter for:

-geo.model <class|object>

Earth model to use for projection. Default: spherical model.

Class Restriction: implements math.geodesy.EarthModel

Default: SphericalVincentyEarthModel

Known implementations:

Parameter for:

-globalpca.filter <class|object>

Filter to use for dimensionality reduction.

Class Restriction: implements math.linearalgebra.pca.filter.EigenPairFilter

Known implementations:

Parameter for:

-globalpca.mode <FULL | CENTER_ROTATE>

Operation mode: full, or rotate only.

Default: FULL

Parameter for:

-glsbs.alpha <double>

Significance niveau

Parameter for:

-glsbs.k <int>

k nearest neighbors to use

Parameter for:

-grid.scale <double>

Scale the grid by the given factor. This can be used to obtain reference points outside the used data space.

Default: 1.0

Parameter for:

-grid.size <int>

The number of partitions in each dimension. Points will be placed on the edges of the grid, except for a grid size of 0, where only the mean is generated as reference point.

Default: 1

Parameter for:

-gridbscan.gridwidth <double>

Width of the grid used, must be at least two times epsilon.

Parameter for:

-hdbscan.hierarchical <|true|false>

Produce a hierarchical output.

Default: false

Parameter for:

-hdbscan.minclsize <int>

The minimum cluster size.

Default: 1

Parameter for:

-hdbscan.minclsize <int>

The minimum cluster size.

Default: 1

Parameter for:

-hdbscan.minPts <int>

Threshold for minimum number of points in the epsilon-neighborhood of a point (including this point).

Parameter for:

-hico.alpha <double>

The threshold for 'strong' eigenvectors: the 'strong' eigenvectors explain a portion of at least alpha of the total variance.

Default: 0.85

Parameter for:

-hico.delta <double>

Threshold of a distance between a vector q and a given space that indicates that q adds a new dimension to the space.

Default: 0.25

Parameter for:

-hico.k <int>

Optional parameter to specify the number of nearest neighbors considered in the PCA. If this parameter is not set, k is set to the value of parameter mu.

Parameter for:

-hico.mu <int>

Specifies the smoothing factor. The mu-nearest neighbor is used to compute the correlation reachability of an object.

Parameter for:

-hics.algo <class|object>

The Algorithm that performs the actual outlier detection on the resulting set of subspace

Class Restriction: implements algorithm.outlier.OutlierAlgorithm

Default: lof.LOF

Known implementations:

Parameter for:

-hics.alpha <double>

The discriminance value that determines the size of the test statistic .

Default: 0.1

Parameter for:

-hics.limit <int>

The threshold that determines how many d-dimensional subspace candidates to retain in each step of the generation

Default: 100

Parameter for:

-hics.m <int>

The number of iterations in the Monte-Carlo processing.

Default: 50

Parameter for:

-hics.seed <long>

The random seed.

Default: use global random seed

Parameter for:

-hics.test <class|object>

The statistical test that is used to calculate the deviation of two data samples

Class Restriction: implements math.statistics.tests.GoodnessOfFitTest

Default: KolmogorovSmirnovTest

Known implementations:

Parameter for:

-hierarchical.hierarchy <|true|false>

Generate a truncated hierarchical clustering result (or strict partitions).

Default: false

Parameter for:

-hierarchical.linkage <class|object>

Linkage method to use (e.g. Ward, Single-Link)

Class Restriction: implements algorithm.clustering.hierarchical.linkage.Linkage

Default: WardLinkage

Known implementations:

Parameter for:

-hierarchical.linkage <SINGLE | COMPLETE | GROUP_AVERAGE | WEIGHTED_AVERAGE | CENTROID | MEDIAN | WARD>

Parameter to choose the linkage strategy.

Default: WARD

Parameter for:

-hierarchical.linkage <SINGLE | COMPLETE | GROUP_AVERAGE | WEIGHTED_AVERAGE | CENTROID | MEDIAN | WARD>

Parameter to choose the linkage strategy.

Default: WARD

Parameter for:

-hierarchical.minclusters <int>

The minimum number of clusters to extract (there may be more clusters when tied, and singletons may be merged into a noise cluster).

Parameter for:

-hierarchical.threshold <double>

The threshold level for which to extract the clusters.

Parameter for:

-HilOut.h <int>

Max. Hilbert-Level

Default: 32

Parameter for:

-HilOut.k <int>

Compute up to k next neighbors

Default: 5

Parameter for:

-HilOut.n <int>

Compute n outliers

Default: 10

Parameter for:

-HilOut.tn <All | TopN>

output of Top n or all elements

Default: TopN

Parameter for:

-hisc.alpha <double>

The maximum absolute variance along a coordinate axis.

Default: 0.01

Parameter for:

-hisc.k <int>

The number of nearest neighbors considered to determine the preference vector. If this value is not defined, k ist set to three times of the dimensionality of the database objects.

Parameter for:

-histogram.splitfreq <|true|false>

Use separate frequencies for outliers and non-outliers.

Default: false

Parameter for:

-holdout.seed <long>

Random generator seed for holdout evaluation.

Default: use global random seed

Parameter for:

-hopkins.k <int>

Nearest neighbor to use for the statistic

Default: 1

Parameter for:

-hopkins.max <double_1,...,double_n>

Maximum values in each dimension. If no value is specified, the maximum value in each dimension will be used. If only one value is specified, this value will be used for all dimensions.

Parameter for:

-hopkins.min <double_1,...,double_n>

Minimum values in each dimension. If no value is specified, the minimum value in each dimension will be used. If only one value is specified, this value will be used for all dimensions.

Parameter for:

-hopkins.rep <int>

The number of times to repeat the experiment (default: 1)

Default: 1

Parameter for:

-hopkins.samplesize <int>

Number of object / random samples to analyze.

Parameter for:

-hopkins.seed <long>

The random number generator.

Default: use global random seed

Parameter for:

-hsbhist.bpp <int_1,...,int_n>

The dimensionality of the histogram in hue, saturation and brightness.

Parameter for:

-hull.alpha <double>

Alpha value for hull drawing (in projected space!).

Default: Infinity

Parameter for:

-id.estimator <class|object>

Class to estimate ID from distance distribution.

Class Restriction: implements math.statistics.intrinsicdimensionality.IntrinsicDimensionalityEstimator

Default: MOMEstimator

Known implementations:

Parameter for:

-id.k <int>

Number of nearest neighbors to use for ID estimation (usually 20-100).

Parameter for:

-idgen.count <int>

Number of DBID to generate.

Parameter for:

-idgen.start <int>

First integer DBID to generate.

Default: 0

Parameter for:

-idist.estimator <class|object>

Estimation method for intrinsic dimensionality.

Class Restriction: implements math.statistics.intrinsicdimensionality.IntrinsicDimensionalityEstimator

Default: GEDEstimator

Known implementations:

Parameter for:

-idist.k <double>

Number of kNN (absolute or relative)

Default: 50.0

Parameter for:

-idist.sampling <double>

Sample size (absolute or relative)

Default: 0.1

Parameter for:

-idistance.distance <class|object>

Distance function to build the index for.

Class Restriction: implements distance.distancefunction.DistanceFunction

Known implementations:

Parameter for:

-idistance.k <int>

Number of reference points to use.

Parameter for:

-idistance.reference <class|object>

Method to choose the reference points.

Class Restriction: implements algorithm.clustering.kmeans.initialization.KMedoidsInitialization

Known implementations:

Parameter for:

-idos.estimator <class|object>

Estimator of intrinsic dimensionality.

Class Restriction: implements math.statistics.intrinsicdimensionality.IntrinsicDimensionalityEstimator

Default: ALIDEstimator

Known implementations:

Parameter for:

-idos.kc <int>

Context set size (ID estimation).

Parameter for:

-idos.kr <int>

Reference set size.

Parameter for:

-index.fill <|true|false>

Partially transparent filling of index pages.

Default: false

Parameter for:

-index.pagefile <class|object>

The pagefile factory for storing the index.

Class Restriction: implements persistent.PageFileFactory

Default: MemoryPageFileFactory

Known implementations:

Parameter for:

-inflo.k <int>

The number of nearest neighbors of an object to be considered for computing its INFLO score.

Parameter for:

-inflo.m <double>

The pruning threshold

Default: 1.0

Parameter for:

-isos.estimator <class|object>

Estimator for intrinsic dimensionality.

Class Restriction: implements math.statistics.intrinsicdimensionality.IntrinsicDimensionalityEstimator

Default: AggregatedHillEstimator

Known implementations:

Parameter for:

-isos.k <int>

Number of neighbors to use. Should be about 3x the desired perplexity.

Default: 100

Parameter for:

-itemsetmining.maxlength <int>

Maximum length of frequent itemsets to report. This can help to reduce the output size to only the most interesting patterns.

Parameter for:

-itemsetmining.minlength <int>

Minimum length of frequent itemsets to report. This can help to reduce the output size to only the most interesting patterns.

Parameter for:

-itemsetmining.minsupp <double>

Threshold for minimum support as minimally required number of transactions (if > 1) or the minimum frequency (if <= 1).

Parameter for:

-itsne.estimator <class|object>

Estimator for intrinsic dimensionality.

Class Restriction: implements math.statistics.intrinsicdimensionality.IntrinsicDimensionalityEstimator

Default: AggregatedHillEstimator

Known implementations:

Parameter for:

-jitter.amount <double>

Jitter amount relative to data.

Parameter for:

-jitter.seed <long>

Jitter random seed.

Default: use global random seed

Parameter for:

-join.sources <object_1|class_1,...,object_n|class_n>

The data sources to join.

Class Restriction: implements datasource.DatabaseConnection

Known implementations:

Parameter for:

-join.sources <object_1|class_1,...,object_n|class_n>

The data sources to join.

Class Restriction: implements datasource.DatabaseConnection

Known implementations:

Parameter for:

-join.sources <object_1|class_1,...,object_n|class_n>

The data sources to join.

Class Restriction: implements datasource.DatabaseConnection

Known implementations:

Parameter for:

-kd.leafsize <int>

Maximum leaf size for the k-d-tree. Nodes will be split until their size is smaller than this threshold.

Default: 1

Parameter for:

-kdeos.idim <int>

Intrinsic dimensionality of this data set. Use -1 for using the true data dimensionality, but values such as 0-2 often offer better performance.

Default: 1

Parameter for:

-kdeos.k.max <int>

Maximum value of k to analyze.

Parameter for:

-kdeos.k.min <int>

Minimum value of k to analyze.

Parameter for:

-kdeos.kernel <class|object>

Kernel density function to use.

Class Restriction: implements math.statistics.kernelfunctions.KernelDensityFunction

Default: GaussianKernelDensityFunction

Known implementations:

Parameter for:

-kdeos.kernel.minbw <double>

Minimum bandwidth for kernel density estimation.

Parameter for:

-kdeos.kernel.scale <double>

Scaling factor for the kernel function.

Default: 0.25

Parameter for:

-kernel.laplace.sigma <double>

Standard deviation of the laplace RBF kernel.

Default: 1.0

Parameter for:

-kernel.polynomial.bias <double>

The bias of the polynomial kernel, a constant that is added to the scalar product.

Parameter for:

-kernel.polynomial.degree <int>

The degree of the polynomial kernel function. Default: 2

Default: 2

Parameter for:

-kernel.rationalquadratic.c <double>

Constant term in the rational quadratic kernel.

Default: 1.0

Parameter for:

-kernel.rbf.sigma <double>

Standard deviation of the Gaussian RBF kernel.

Default: 1.0

Parameter for:

-kernel.sigmoid.c <double>

Sigmoid c parameter (scaling).

Default: 1.0

Parameter for:

-kernel.sigmoid.theta <double>

Sigmoid theta parameter (bias).

Default: 0.0

Parameter for:

-kernelcluster.dim <int>

Dimension to use for clustering. For one-dimensional data, use 0.

Default: 0

Parameter for:

-kernelcluster.kernel <class|object>

Kernel function for density estimation.

Class Restriction: implements math.statistics.kernelfunctions.KernelDensityFunction

Default: EpanechnikovKernelDensityFunction

Known implementations:

Parameter for:

-kernelcluster.knn <int>

Number of nearest neighbors to use for bandwidth estimation.

Parameter for:

-kernelcluster.mode <BALLOON | SAMPLE>

Kernel density estimation mode (baloon estimator vs. sample point estimator).

Default: BALLOON

Parameter for:

-kernelcluster.window <int>

Half width of sliding window to find local minima.

Parameter for:

-kernellof.kernel <class|object>

Kernel to use for kernel density LOF.

Class Restriction: implements math.statistics.kernelfunctions.KernelDensityFunction

Default: EpanechnikovKernelDensityFunction

Known implementations:

Parameter for:

-kmeans.algorithm <class|object>

KMeans variant to run multiple times.

Class Restriction: implements algorithm.clustering.kmeans.KMeans

Known implementations:

Parameter for:

-kmeans.algorithm <class|object>

Clustering algorithm to use for detecting outliers.

Class Restriction: implements algorithm.clustering.kmeans.KMeans

Default: KMeansLloyd

Known implementations:

Parameter for:

-kmeans.algorithm <class|object>

KMeans variant to run multiple times.

Class Restriction: implements algorithm.clustering.kmeans.KMeans

Known implementations:

Parameter for:

-kmeans.initialization <class|object>

Method to choose the initial means.

Class Restriction: implements algorithm.clustering.kmeans.initialization.KMedoidsInitialization

Default: PAMInitialMeans

Known implementations:

Parameter for:

-kmeans.k <int>

The number of clusters to find.

Parameter for:

-kmeans.maxiter <int>

The maximum number of iterations to do. 0 means no limit.

Parameter for:

-kmeans.means <double_11,...,double_1n:...:double_m1,...,double_mn>

Initial means for k-means.

Parameter for:

-kmeans.qualitymeasure <class|object>

Quality measure variant for deciding which run to keep.

Class Restriction: implements algorithm.clustering.kmeans.quality.KMeansQualityMeasure

Known implementations:

Parameter for:

-kmeans.samplesize <double>

Sample set size (if > 1) or sampling rante (if < 1).

Parameter for:

-kmeans.seed <long>

The random number generator seed.

Default: use global random seed

Parameter for:

-kmeans.trials <int>

The number of trials to run.

Parameter for:

-kmeans.varstat <|true|false>

Compute the final clustering variance statistic. Needs an additional full pass over the data set.

Default: false

Parameter for:

-kmeansmm.l <double>

Number of outliers to ignore, or (if less than 1) a relative rate.

Default: 0.05

Parameter for:

-kmeansmm.noisecluster <|true|false>

Create a noise cluster, instead of assigning the noise objects.

Default: false

Parameter for:

-kml.autoopen <|true|false>

Automatically open the result file.

Default: false

Parameter for:

-kml.compat <|true|false>

Use simpler KML objects, compatibility mode.

Default: false

Parameter for:

-kml.scaling <class|object>

Additional scaling function for KML colorization.

Class Restriction: implements utilities.scaling.outlier.OutlierScaling

Default: OutlierLinearScaling

Known implementations:

Parameter for:

-knnbench.k <int>

Number of neighbors to retreive for kNN benchmarking.

Parameter for:

-knnbench.query <class|object>

Data source for the queries. If not set, the queries are taken from the database.

Class Restriction: implements datasource.DatabaseConnection

Known implementations:

Parameter for:

-knnbench.random <long>

Random generator for sampling.

Default: use global random seed

Parameter for:

-knnbench.sampling <double>

Sampling size parameter. If the value is less or equal 1, it is assumed to be the relative share. Larger values will be interpreted as integer sizes. By default, all data will be used.

Parameter for:

-knnclassifier.k <int>

The number of neighbors to take into account for classification.

Default: 1

Parameter for:

-knndd.k <int>

The k nearest neighbor, excluding the query point (i.e. query point is the 0-nearest-neighbor)

Default: 1

Parameter for:

-knndistanceorder.k <int>

Specifies the distance of the k-distant object to be assessed, ignoring the query object.

Parameter for:

-knndistanceorder.sample <double>

The percentage of objects to use for sampling, or the absolute number of samples.

Default: 1.0

Parameter for:

-knndistanceorder.seed <long>

Random generator seed for sampling.

Default: use global random seed

Parameter for:

-knngraph.delta <double>

The early termination parameter.

Default: 0.001

Parameter for:

-knngraph.maxiter <int>

maximum number of iterations

Default: 100

Parameter for:

-knngraph.no-initial <|true|false>

Do not use initial neighbors.

Default: false

Parameter for:

-knngraph.rho <double>

The sample rate parameter

Default: 1.0

Parameter for:

-knngraph.seed <long>

The random number seed.

Default: use global random seed

Parameter for:

-knnjoin.k <int>

Specifies the k-nearest neighbors to be assigned.

Default: 1

Parameter for:

-knno.k <int>

The k nearest neighbor, excluding the query point (i.e. query point is the 0-nearest-neighbor)

Parameter for:

-knnwod.k <int>

The k nearest neighbor, excluding the query point (i.e. query point is the 0-nearest-neighbor)

Parameter for:

-lancewilliams.beta <double>

Beta for the Lance-Williams flexible beta approach.

Default: -0.25

Parameter for:

-lcss.pDelta <double>

the allowed deviation in x direction for LCSS alignment (positive double value, 0 <= pDelta <= 1)

Default: 0.1

Parameter for:

-lcss.pEpsilon <double>

the allowed deviation in y direction for LCSS alignment (positive double value, 0 <= pEpsilon <= 1)

Default: 0.05

Parameter for:

-ldf.c <double>

Score scaling parameter for LDF.

Default: 0.1

Parameter for:

-ldf.h <double>

Kernel bandwidth multiplier for LDF.

Parameter for:

-ldf.k <int>

Number of neighbors to use for LDF.

Parameter for:

-ldf.kernel <class|object>

Kernel to use for LDF.

Class Restriction: implements math.statistics.kernelfunctions.KernelDensityFunction

Default: GaussianKernelDensityFunction

Known implementations:

Parameter for:

-ldof.k <int>

The number of nearest neighbors of an object to be considered for computing its LDOF_SCORE.

Parameter for:

-leader.threshold <double>

Maximum distance from leading object.

Parameter for:

-lic.k <int>

The k nearest neighbor, excluding the query point (i.e. query point is the 0-nearest-neighbor)

Parameter for:

-linearscale.ignorezero <|true|false>

Ignore zero entries when computing the minimum and maximum.

Default: false

Parameter for:

-linearscale.max <double>

Fixed maximum to use in linear scaling.

Parameter for:

-linearscale.min <double>

Fixed minimum to use in linear scaling.

Parameter for:

-linearscale.usemean <|true|false>

Use the mean as minimum for scaling.

Default: false

Parameter for:

-lmclus.maxdim <int>

Maximum linear manifold dimension to search.

Parameter for:

-lmclus.minsize <int>

Minimum cluster size to allow.

Parameter for:

-lmclus.sampling-level <int>

A number used to determine how many samples are taken in each search.

Default: 100

Parameter for:

-lmclus.seed <long>

Random generator seed.

Default: use global random seed

Parameter for:

-lmclus.threshold <double>

Threshold to determine if a cluster was found.

Parameter for:

-localpca.distancefunction <class|object>

The distance function used to select objects for running PCA.

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-localpca.k <int>

The number of nearest neighbors considered in the PCA. If this parameter is not set, k ist set to three times of the dimensionality of the database objects.

Parameter for:

-loci.alpha <int>

Scaling factor for averaging neighborhood

Default: 4

Parameter for:

-loci.alpha <double>

Scaling factor for averaging neighborhood

Default: 0.5

Parameter for:

-loci.g <int>

The number of Grids to use.

Default: 1

Parameter for:

-loci.nmin <int>

Minimum neighborhood size to be considered.

Default: 20

Parameter for:

-loci.nmin <int>

Minimum neighborhood size to be considered.

Default: 20

Parameter for:

-loci.rmax <double>

The maximum radius of the neighborhood to be considered.

Parameter for:

-loci.seed <long>

The seed to use for initializing Random.

Default: use global random seed

Parameter for:

-lof.k <int>

The number of nearest neighbors (not including the query point) of an object to be considered for computing its LOF score.

Parameter for:

-lof.kreach <int>

The number of nearest neighbors of an object to be considered for computing its LOF score.

Parameter for:

-lof.krefer <int>

The number of nearest neighbors of an object to be considered for computing its LOF score.

Parameter for:

-lof.reachdistfunction <class|object>

Distance function to determine the reachability distance between database objects.

Class Restriction: implements distance.distancefunction.DistanceFunction

Known implementations:

Parameter for:

-log1pscale.boost <double>

Boosting factor. Larger values will yield a steeper curve.

Default: 1.0

Parameter for:

-loop.comparedistfunction <class|object>

Distance function to determine the reference set of an object.

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-loop.kcomp <int>

The number of nearest neighbors of an object to be considered for computing its LOOP_SCORE.

Parameter for:

-loop.kref <int>

The number of nearest neighbors of an object to be used for the PRD value.

Parameter for:

-loop.lambda <double>

The number of standard deviations to consider for density computation.

Default: 2.0

Parameter for:

-loop.referencedistfunction <class|object>

Distance function to determine the density of an object.

Class Restriction: implements distance.distancefunction.DistanceFunction

Known implementations:

Parameter for:

-lpnorm.p <int>

Degree p of the L_p-Norm (positive number)

Parameter for:

-lsdbc.alpha <double>

Density difference factor

Parameter for:

-lsdbc.k <int>

Neighborhood size (k)

Parameter for:

-lsh.buckets <int>

Number of hash buckets to use.

Default: 7919

Parameter for:

-lsh.family <class|object>

Hash function family to use for LSH.

Class Restriction: implements index.lsh.hashfamilies.LocalitySensitiveHashFunctionFamily

Known implementations:

Parameter for:

-lsh.projection.projections <int>

Number of projections to use for each hash function.

Parameter for:

-lsh.projection.projections <int>

Number of projections to use for each hash function.

Parameter for:

-lsh.projection.random <long>

Random seed for generating the projections.

Default: use global random seed

Parameter for:

-lsh.projection.random <long>

Random seed for generating the projections.

Default: use global random seed

Parameter for:

-lsh.projection.width <double>

Bin width for random projections.

Parameter for:

-lsh.tables <int>

Number of hash tables to use.

Parameter for:

-map.includeself <|true|false>

Include the query object in the evaluation.

Default: false

Parameter for:

-map.maxk <int>

Maximum value of k for kNN evaluation.

Parameter for:

-map.sampling <double>

Relative amount of object to sample.

Parameter for:

-map.sampling-seed <long>

Random seed for deterministic sampling.

Default: use global random seed

Parameter for:

-materialize.distance <class|object>

the distance function to materialize the nearest neighbors

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-materialize.k <int>

The number of nearest neighbors of an object to be materialized.

Parameter for:

-matrix.distance <class|object>

Distance function for the precomputed distance matrix.

Class Restriction: implements distance.distancefunction.DistanceFunction

Known implementations:

Parameter for:

-matrix.similarity <class|object>

Similarity function for the precomputed similarity matrix.

Class Restriction: implements distance.similarityfunction.SimilarityFunction

Known implementations:

Parameter for:

-mds.dim <int>

Output dimensionality.

Parameter for:

-mds.distance <class|object>

Distance function to use.

Class Restriction: implements distance.distancefunction.PrimitiveDistanceFunction

Default: minkowski.SquaredEuclideanDistanceFunction

Known implementations:

Parameter for:

-mds.seed <long>

Random seed for fast MDS.

Default: use global random seed

Parameter for:

-mds.vector-type <class|object>

The type of vectors to create.

Class Restriction: implements data.NumberVector

Default: DoubleVector

Known implementations:

Parameter for:

-meanshift.kernel <class|object>

Kernel function to use with mean-shift clustering.

Class Restriction: implements math.statistics.kernelfunctions.KernelDensityFunction

Default: EpanechnikovKernelDensityFunction

Known implementations:

Parameter for:

-meanshift.kernel-bandwidth <double>

Range of the kernel to use (aka: radius, bandwidth).

Parameter for:

-metaoutlier.scaling <class|object>

Class to use as scaling function.

Class Restriction: implements utilities.scaling.ScalingFunction

Known implementations:

Parameter for:

-mkapp.k <int>

positive integer specifying the maximum number k of reverse k nearest neighbors to be supported.

Parameter for:

-mkapp.nolog <|true|false>

Flag to indicate that the approximation is done in the ''normal'' space instead of the log-log space (which is default).

Default: false

Parameter for:

-mkapp.p <int>

positive integer specifying the order of the polynomial approximation.

Parameter for:

-mkcop.k <int>

positive integer specifying the maximum number k of reverse k nearest neighbors to be supported.

Parameter for:

-mktree.kmax <int>

Specifies the maximal number k of reverse k nearest neighbors to be supported.

Parameter for:

-mmo.c <double>

cutoff

Default: 1.0E-7

Parameter for:

-modeloutlier.expect <double>

Expected amount of outliers, for making the scores more intuitive. When the value is 1, the CDF will be given instead.

Default: 0.01

Parameter for:

-mtree.distancefunction <class|object>

Distance function to determine the distance between database objects.

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-mtree.insert <class|object>

Insertion strategy to use for constructing the M-tree.

Class Restriction: implements index.tree.metrical.mtreevariants.strategies.insert.MTreeInsert

Default: MinimumEnlargementInsert

Known implementations:

Parameter for:

-mtree.randomsplit.random <long>

Random generator / seed for the randomized split.

Default: use global random seed

Parameter for:

-mtree.split <class|object>

Split strategy to use for constructing the M-tree.

Class Restriction: implements index.tree.metrical.mtreevariants.strategies.split.MTreeSplit

Default: MLBDistSplit

Known implementations:

Parameter for:

-mtree.split.distributor <class|object>

Distribution strategy for mtree entries during splitting.

Class Restriction: implements index.tree.metrical.mtreevariants.strategies.split.distribution.DistributionStrategy

Default: GeneralizedHyperplaneDistribution

Known implementations:

Parameter for:

-multinorm.ps <double_1,...,double_n>

The exponents to use for this distance function

Parameter for:

-nanfilter.replacement <class|object>

Distribution to sample replacement values from.

Class Restriction: implements math.statistics.distribution.Distribution

Known implementations:

Parameter for:

-neighborhood <class|object>

The neighborhood predicate to use in comparison step.

Class Restriction: implements algorithm.outlier.spatial.neighborhood.NeighborSetPredicate

Known implementations:

Parameter for:

-neighborhood.distancefunction <class|object>

the distance function to use

Class Restriction: implements distance.distancefunction.DistanceFunction

Known implementations:

Parameter for:

-neighborhood.inner <class|object>

Parameter for the non-weighted neighborhood to use.

Class Restriction: implements algorithm.outlier.spatial.neighborhood.NeighborSetPredicate

Known implementations:

Parameter for:

-neighborhood.k <int>

the number of neighbors

Parameter for:

-nfold <int>

positive number of folds for cross-validation

Default: 10

Parameter for:

-nfold <int>

Number of folds for cross-validation.

Default: 10

Parameter for:

-nfold <int>

Number of folds for cross-validation

Default: 10

Parameter for:

-normalization.max <double>

Maximum value to assign to objects.

Default: 1.0

Parameter for:

-normalization.min <double>

Minimum value to assign to objects.

Default: 0.0

Parameter for:

-normalization.norm <class|object>

Norm (length function) to use for computing the vector length.

Class Restriction: implements distance.distancefunction.Norm

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-normalize.beta.alpha <double>

Alpha parameter to control the shape of the output distribution.

Default: 0.1

Parameter for:

-normalize.distributions <object_1|class_1,...,object_n|class_n>

A list of the distribution estimators to try.

Class Restriction: implements math.statistics.distribution.estimator.DistributionEstimator

Default: meta.BestFitEstimator

Known implementations:

Parameter for:

-normalize.max <double_1,...,double_n>

a comma separated concatenation of the maximum values in each dimension that are mapped to 1. If no value is specified, the maximum value of the attribute range in this dimension will be taken.

Parameter for:

-normalize.mean <double_1,...,double_n>

a comma separated concatenation of the mean values in each dimension that are mapped to 0. If no value is specified, the mean value of the attribute range in this dimension will be taken.

Parameter for:

-normalize.min <double_1,...,double_n>

a comma separated concatenation of the minimum values in each dimension that are mapped to 0. If no value is specified, the minimum value of the attribute range in this dimension will be taken.

Parameter for:

-normalize.stddev <double_1,...,double_n>

a comma separated concatenation of the standard deviations in each dimension that are scaled to 1. If no value is specified, the standard deviation of the attribute range in this dimension will be taken.

Parameter for:

-odin.k <int>

Number of neighbors to use for kNN graph.

Parameter for:

-odin.k <int>

Number of neighbors to use for kNN graph.

Parameter for:

-optics.epsilon <double>

The maximum radius of the neighborhood to be considered.

Parameter for:

-optics.minpts <int>

Threshold for minimum number of points in the epsilon-neighborhood of a point.

Parameter for:

-opticsxi.algorithm <class>

The actual OPTICS-type algorithm to use.

Class Restriction: implements algorithm.clustering.optics.OPTICSTypeAlgorithm

Default: OPTICSHeap

Known implementations:

Parameter for:

-opticsxi.keepsteep <|true|false>

Keep the steep up/down areas of the plot.

Default: false

Parameter for:

-opticsxi.nocorrect <|true|false>

Disable the predecessor correction.

Default: false

Parameter for:

-opticsxi.xi <double>

Threshold for the steepness requirement.

Parameter for:

-orclus.alpha <double>

The factor for reducing the number of current clusters in each iteration.

Default: 0.5

Parameter for:

-orclus.seed <long>

The random number generator seed.

Default: use global random seed

Parameter for:

-out <file>

Filename the KMZ file (compressed KML) is written to.

Parameter for:

-out.filter <pattern>

Filter pattern for output selection. Only output streams that match the given pattern will be written.

Parameter for:

-out.gzip <|true|false>

Enable gzip compression of output files.

Default: false

Parameter for:

-out.silentoverwrite <|true|false>

Silently overwrite output files.

Default: false

Parameter for:

-outlier.pattern <pattern>

Label pattern to match outliers.

Default: .*(Outlier|Noise).*

Parameter for:

-outliereval.positive <pattern>

Class label for the 'positive' class.

Parameter for:

-outrank.algorithm <class|object>

Subspace clustering algorithm to use.

Class Restriction: implements algorithm.clustering.subspace.SubspaceClusteringAlgorithm

Known implementations:

Parameter for:

-outrank.s1.alpha <double>

Alpha parameter for S1 score.

Default: 0.25

Parameter for:

-outres.epsilon <double>

Range value for OUTRES in 2 dimensions.

Parameter for:

-p3c.alpha <double>

The significance level for uniform testing in the initial binning step.

Default: 0.001

Parameter for:

-p3c.em.delta <double>

The change delta for the EM step below which to stop.

Default: 1.0E-5

Parameter for:

-p3c.em.maxiter <int>

The maximum number of iterations for the EM step. Use -1 to run until delta convergence.

Default: 20

Parameter for:

-p3c.minsize <int>

The minimum size of a cluster, otherwise it is seen as noise (this is a cheat, it is not mentioned in the paper).

Default: 1

Parameter for:

-p3c.threshold <double>

The threshold value for the poisson test used when merging signatures.

Default: 1.0E-4

Parameter for:

-pagefile.cachesize <int>

The size of the cache in bytes.

Parameter for:

-pagefile.file <file>

The name of the file storing the page file.

Parameter for:

-pagefile.file <file>

The name of the file storing the page file.

Parameter for:

-pagefile.pagefile <class|object>

The backing pagefile for the cache.

Class Restriction: implements persistent.PageFileFactory

Default: PersistentPageFileFactory

Known implementations:

Parameter for:

-pagefile.pagesize <int>

The size of a page in bytes.

Default: 1024

Parameter for:

-paircounting.noisespecial <|true|false>

Use special handling for noise clusters.

Default: false

Parameter for:

-paircounting.reference <class|object>

Reference clustering to compare with. Defaults to a by-label clustering.

Class Restriction: implements algorithm.clustering.ClusteringAlgorithm

Default: trivial.ByLabelOrAllInOneClustering

Known implementations:

Parameter for:

-paircounting.selfpair <|true|false>

Enable self-pairing for cluster comparison.

Default: false

Parameter for:

-pam.fasttol <double>

Tolerance for optimistically performing additional swaps, where 1 executes all fast swaps, 0 only those that are unaffected by the primary swaps.

Default: 1.0

Parameter for:

-parallel.clusteroutline.straight <|true|false>

Draw straight lines

Default: false

Parameter for:

-parallel3d.layout <class|object>

Layouting method for 3DPC.

Class Restriction: implements visualization.parallel3d.layout.Layouter3DPC

Default: SimpleCircularMSTLayout3DPC

Known implementations:

Parameter for:

-parallel3d.sim <class|object>

Similarity measure for spanning tree.

Class Restriction: implements math.statistics.dependence.DependenceMeasure

Default: CorrelationDependenceMeasure

Known implementations:

Parameter for:

-parser.colsep <pattern>

Column separator pattern. The default assumes whitespace separated data.

Default: \s*[,;\s]\s*

Parameter for:

-parser.labelIndices <int_1,...,int_n>

A comma separated list of the indices of labels (may be numeric), counting whitespace separated entries in a line starting with 0. The corresponding entries will be treated as a label.

Parameter for:

-parser.quote <string>

Quotation characters. By default, both double and single ASCII quotes are accepted.

Default: "'

Parameter for:

-parser.vector-type <class|object>

The type of vectors to create for numerical attributes.

Class Restriction: implements data.NumberVector

Default: DoubleVector

Known implementations:

Parameter for:

-partknn.p <int>

The number of partitions to use for approximate kNN.

Parameter for:

-partknn.seed <long>

The random number generator seed.

Default: use global random seed

Parameter for:

-patternfilter.invert <|true|false>

Flag to invert pattern.

Default: false

Parameter for:

-patternfilter.pattern <pattern>

The filter pattern to use.

Parameter for:

-pbm.distance <class|object>

Distance function to use for computing PBM.

Class Restriction: implements distance.distancefunction.NumberVectorDistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-pbm.noisehandling <MERGE_NOISE | TREAT_NOISE_AS_SINGLETONS | IGNORE_NOISE>

Control how noise should be treated.

Default: TREAT_NOISE_AS_SINGLETONS

Parameter for:

-pca.covariance <class|object>

Class used to compute the covariance matrix.

Class Restriction: implements math.linearalgebra.pca.CovarianceMatrixBuilder

Default: StandardCovarianceMatrixBuilder

Known implementations:

Parameter for:

-pca.filter <class|object>

Filter class to determine the strong and weak eigenvectors.

Class Restriction: implements math.linearalgebra.pca.filter.EigenPairFilter

Default: PercentageEigenPairFilter

Known implementations:

Parameter for:

-pca.filter.absolute <|true|false>

Flag to mark delta as an absolute value.

Default: false

Parameter for:

-pca.filter.alpha <double>

The share (0.0 to 1.0) of variance that needs to be explained by the 'strong' eigenvectors. The filter class will choose the number of strong eigenvectors by this share.

Default: 0.85

Parameter for:

-pca.filter.delta <double>

The threshold for strong Eigenvalues. If not otherwise specified, delta is a relative value w.r.t. the (absolute) highest Eigenvalues and has to be a double between 0 and 1. To mark delta as an absolute value, use the option -pca.filter.absolute.

Default: 0.1

Parameter for:

-pca.filter.n <int>

The number of strong eigenvectors: n eigenvectors with the n highest eigenvalues are marked as strong eigenvectors.

Parameter for:

-pca.filter.progressivealpha <double>

The share (0.0 to 1.0) of variance that needs to be explained by the 'strong' eigenvectors. The filter class will choose the number of strong eigenvectors by this share.

Default: 0.5

Parameter for:

-pca.filter.relativealpha <double>

The sensitivity niveau for weak eigenvectors: An eigenvector which is at less than the given share of the statistical average variance is considered weak.

Default: 1.1

Parameter for:

-pca.filter.weakalpha <double>

The minimum strength of the statistically expected variance (1/n) share an eigenvector needs to have to be considered 'strong'.

Default: 0.0

Parameter for:

-pca.variant <class|object>

The class to compute (filtered) PCA.

Class Restriction: extends math.linearalgebra.pca.PCARunner

Default: PCARunner

Known implementations:

Parameter for:

-pca.weight <class|object>

Weight function to use in weighted PCA.

Class Restriction: implements math.linearalgebra.pca.weightfunctions.WeightFunction

Default: ConstantWeight

Known implementations:

Parameter for:

-perturbationfilter.max <double_1,...,double_n>

Only used, if MINMAX is set as scaling reference: a comma separated concatenation of the maximum values in each dimension assumed as a reference. If no value is specified, the maximum value of the attribute range in this dimension will be taken.

Parameter for:

-perturbationfilter.min <double_1,...,double_n>

Only used, if MINMAX is set as scaling reference: a comma separated concatenation of the minimum values in each dimension assumed as a reference. If no value is specified, the minimum value of the attribute range in this dimension will be taken.

Parameter for:

-perturbationfilter.noisedistribution <GAUSSIAN | UNIFORM>

The nature of the noise distribution, default is UNIFORM

Default: UNIFORM

Parameter for:

-perturbationfilter.percentage <double>

Percentage of the standard deviation of the random Gaussian noise generation per attribute, given the standard deviation of the corresponding attribute in the original data distribution (assuming a Gaussian distribution there).

Default: 0.01

Parameter for:

-perturbationfilter.scalingreference <UNITCUBE | STDDEV | MINMAX>

The reference for scaling the Gaussian noise. Default is UNITCUBE, parameter perturbationfilter.percentage will then directly define the standard deviation of all noise Gaussians. For options STDDEV and MINMAX, the percentage of the attributewise standard deviation or extension, repectively, will define the attributewise standard deviation of the noise Gaussians.

Default: UNITCUBE

Parameter for:

-perturbationfilter.seed <long>

Seed for random noise generation.

Parameter for:

-pinn.hmult <double>

Multiplicator for neighborhood size.

Default: 3.0

Parameter for:

-pinn.s <double>

Sparsity of the random projection.

Default: 1.0

Parameter for:

-pinn.seed <long>

Random generator seed.

Default: use global random seed

Parameter for:

-pinn.t <int>

Target dimensionality.

Parameter for:

-precision.k <int>

k value for precision@k. Can be set to 0, to get R-precision, or the precision-recall-break-even-point.

Default: 0

Parameter for:

-precision.maxk <int>

Maximum value of 'k' to compute the curve up to.

Parameter for:

-precision.positive <pattern>

Class label for the 'positive' class.

Parameter for:

-precision.positive <pattern>

Class label for the 'positive' class.

Parameter for:

-predecon.delta <double>

A double specifying the variance threshold for small Eigenvalues.

Parameter for:

-predecon.kappa <double>

Penalty factor for deviations in preferred (low-variance) dimensions.

Default: 20.0

Parameter for:

-predecon.kappa <double>

Penalty factor for deviations in preferred (low-variance) dimensions.

Default: 20.0

Parameter for:

-predecon.lambda <int>

Maximum dimensionality to consider for core points.

Parameter for:

-predecon.lambda <int>

Maximum dimensionality to consider for core points.

Parameter for:

-proclus.mi <int>

The multiplier for the initial number of medoids.

Default: 10

Parameter for:

-proclus.seed <long>

The random number generator seed.

Default: use global random seed

Parameter for:

-projectedclustering.k <int>

The number of clusters to find.

Parameter for:

-projectedclustering.k_i <int>

The multiplier for the initial number of seeds.

Default: 30

Parameter for:

-projectedclustering.l <int>

The dimensionality of the clusters to find.

Parameter for:

-projection <class|object>

Projection to use.

Class Restriction: implements data.projection.Projection

Known implementations:

Parameter for:

-projection.dim <int>

Projection dimensionality

Default: 2

Parameter for:

-projectionfilter.selectedattributes <int_1,...,int_n>

a comma separated array of integer values d_i, where 0 <= d_i < the dimensionality of the feature space specifying the dimensions to be considered for projection. If this parameter is not set, no dimensions will be considered, i.e. the projection is a zero-dimensional feature space

Parameter for:

-projections.family <class|object>

Random projection family to use. The default is to use the original axes.

Class Restriction: implements data.projection.random.RandomProjectionFamily

Known implementations:

Parameter for:

-projections.projections <int>

Number of projections to use.

Parameter for:

-projections.seed <long>

Random generator.

Default: use global random seed

Parameter for:

-projections.windowmult <double>

Window size multiplicator.

Default: 10.0

Parameter for:

-projhistogram.bins <int>

Number of bins in the distribution histogram

Default: 80

Parameter for:

-projhistogram.curves <|true|false>

Use curves instead of the stacked histogram style.

Default: false

Parameter for:

-projindex.disable-refine <|true|false>

Flag to disable refinement of distances.

Default: false

Parameter for:

-projindex.inner <class|object>

Index to use on the projected data.

Class Restriction: implements index.IndexFactory

Known implementations:

Parameter for:

-projindex.kmulti <double>

Multiplier for k.

Default: 1.0

Parameter for:

-projindex.materialize <|true|false>

Flag to materialize the projected data.

Default: false

Parameter for:

-projindex.proj <class|object>

Projection to use for the projected index.

Class Restriction: implements data.projection.Projection

Known implementations:

Parameter for:

-pwc.alpha <double>

Alpha threshold for estimating the confidence probability.

Default: 0.95

Parameter for:

-pwc.clustering <class|object>

Clustering algorithm used on the samples.

Class Restriction: implements algorithm.clustering.ClusteringAlgorithm

Known implementations:

Parameter for:

-pwc.distance <class|object>

Distance measure of clusterings.

Class Restriction: implements distance.similarityfunction.cluster.ClusteringDistanceSimilarityFunction

Default: ClusteringAdjustedRandIndexSimilarityFunction

Known implementations:

Parameter for:

-pwc.metaclustering <class|object>

Algorithm used to aggregate clustering results. Must be a distance-based clustering algorithm.

Class Restriction: implements algorithm.clustering.ClusteringAlgorithm

Default: kmeans.KMedoidsPAM

Known implementations:

Parameter for:

-pwc.random <long>

Random generator used for sampling.

Default: use global random seed

Parameter for:

-pwc.samples <int>

Number of clusterings to produce on samples.

Default: 10

Parameter for:

-pwc.samples.keep <|true|false>

Retain all sampled relations, not only the representative results.

Default: false

Parameter for:

-randomknn.seed <long>

The random number seed.

Default: use global random seed

Parameter for:

-randomknn.share <double>

The relative amount of objects to consider for kNN computations.

Parameter for:

-randomproj.dimensionality <int>

Amount of dimensions to project to.

Parameter for:

-randomproj.family <class|object>

Projection family to use.

Class Restriction: implements data.projection.random.RandomProjectionFamily

Default: AchlioptasRandomProjectionFamily

Known implementations:

Parameter for:

-randomproj.random <long>

Random generator seed.

Default: use global random seed

Parameter for:

-randomprojection.numberselected <int>

number of selected attributes

Default: 1

Parameter for:

-randomprojection.seed <long>

Seed for random selection of projection attributes.

Default: use global random seed

Parameter for:

-randomwalkec.alpha <double>

Scaling exponent for value differences.

Default: 0.5

Parameter for:

-randomwalkec.c <double>

The damping parameter c.

Parameter for:

-randomwalkec.k <int>

Number of nearest neighbors to use.

Parameter for:

-rangebench.query <class|object>

Data source for the queries. If not set, the queries are taken from the database.

Class Restriction: implements datasource.DatabaseConnection

Known implementations:

Parameter for:

-rangebench.random <long>

Random generator for sampling.

Default: use global random seed

Parameter for:

-rangebench.sampling <double>

Sampling size parameter. If the value is less or equal 1, it is assumed to be the relative share. Larger values will be interpreted as integer sizes. By default, all data will be used.

Parameter for:

-rankqual.bins <int>

Number of bins to use in the histogram

Default: 20

Parameter for:

-rankqual.bins <int>

Number of bins to use in the histogram

Default: 100

Parameter for:

-ransacpca.iterations <int>

The number of iterations to perform.

Default: 1000

Parameter for:

-ransacpca.seed <long>

Random seed (optional).

Default: use global random seed

Parameter for:

-rdknn.distancefunction <class|object>

Distance function to determine the distance between database objects.

Class Restriction: implements distance.distancefunction.SpatialPrimitiveDistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-rdknn.k <int>

positive integer specifying the maximal number k of reverse k nearest neighbors to be supported.

Parameter for:

-refod.k <int>

The number of nearest neighbors

Parameter for:

-refod.refp <class|object>

The heuristic for finding reference points.

Class Restriction: implements utilities.referencepoints.ReferencePointsHeuristic

Default: GridBasedReferencePoints

Known implementations:

Parameter for:

-resulthandler <object_1|class_1,...,object_n|class_n>

Result handler class.

Class Restriction: implements result.ResultHandler

Known implementations:

Parameter for:

-rgbhist.bpp <int>

The dimensionality of the histogram in each color

Parameter for:

-rocauc.positive <pattern>

Class label for the 'positive' class.

Parameter for:

-rtree.bulk.spatial-sort <class|object>

Strategy for spatial sorting in bulk loading.

Class Restriction: implements math.spacefillingcurves.SpatialSorter

Known implementations:

Parameter for:

-rtree.insert-directory <class>

Insertion strategy for directory nodes.

Class Restriction: implements index.tree.spatial.rstarvariants.strategies.insert.InsertionStrategy

Default: LeastEnlargementWithAreaInsertionStrategy

Known implementations:

Parameter for:

-rtree.insert-leaf <class>

Insertion strategy for leaf nodes.

Class Restriction: implements index.tree.spatial.rstarvariants.strategies.insert.InsertionStrategy

Default: LeastOverlapInsertionStrategy

Known implementations:

Parameter for:

-rtree.insertion-candidates <int>

defines how many children are tested for finding the child generating the least overlap when inserting an object.

Default: 32

Parameter for:

-rtree.insertionstrategy <class|object>

The strategy to use for object insertion.

Class Restriction: implements index.tree.spatial.rstarvariants.strategies.insert.InsertionStrategy

Default: CombinedInsertionStrategy

Known implementations:

Parameter for:

-rtree.minimum-fill <double>

Minimum relative fill required for data pages.

Default: 0.4

Parameter for:

-rtree.overflowtreatment <class|object>

The strategy to use for handling overflows.

Class Restriction: implements index.tree.spatial.rstarvariants.strategies.overflow.OverflowTreatment

Default: LimitedReinsertOverflowTreatment

Known implementations:

Parameter for:

-rtree.reinsertion-amount <double>

The amount of entries to reinsert.

Default: 0.3

Parameter for:

-rtree.reinsertion-distancce <class|object>

The distance function to compute reinsertion candidates by.

Class Restriction: implements distance.distancefunction.SpatialPrimitiveDistanceFunction

Default: minkowski.SquaredEuclideanDistanceFunction

Known implementations:

Parameter for:

-rtree.reinsertion-strategy <class|object>

The strategy to select candidates for reinsertion.

Class Restriction: implements index.tree.spatial.rstarvariants.strategies.reinsert.ReinsertStrategy

Default: CloseReinsert

Known implementations:

Parameter for:

-rtree.splitstrategy <class|object>

The strategy to use for node splitting.

Class Restriction: implements index.tree.spatial.rstarvariants.strategies.split.SplitStrategy

Default: TopologicalSplitter

Known implementations:

Parameter for:

-sample.n <int>

The number of samples to draw.

Parameter for:

-sample.random <long>

Random generator seed.

Default: use global random seed

Parameter for:

-sampling.p <double>

Sampling probability. Each object has a chance of being sampled with this probability.

Parameter for:

-sampling.seed <long>

Random generator seed for sampling.

Default: use global random seed

Parameter for:

-scales.minmax <double_1,...,double_n>

Forcibly set the scales to the given range.

Parameter for:

-scaling.gamma <double>

Gamma value for scaling.

Parameter for:

-selectionrange.nofill <|true|false>

Use wireframe style for selection ranges.

Default: false

Parameter for:

-selectivity.radius <double>

Radius to use for selectivity estimation.

Parameter for:

-selectivity.sampling <double>

Relative amount of object to sample.

Parameter for:

-selectivity.sampling-seed <long>

Random seed for deterministic sampling.

Default: use global random seed

Parameter for:

-series.variates <int>

Number of variates this time series has.

Parameter for:

-sfcknn.curves <object_1|class_1,...,object_n|class_n>

Space filling curve generators to use for kNN approximation.

Class Restriction: implements math.spacefillingcurves.SpatialSorter

Known implementations:

Parameter for:

-sfcknn.curves <object_1|class_1,...,object_n|class_n>

Space filling curve generators to use for kNN approximation.

Class Restriction: implements math.spacefillingcurves.SpatialSorter

Known implementations:

Parameter for:

-sfcknn.dim <int>

Number of dimensions to use for each curve.

Parameter for:

-sfcknn.proj <class|object>

Random projection to use.

Class Restriction: implements data.projection.random.RandomProjectionFamily

Known implementations:

Parameter for:

-sfcknn.seed <long>

Random generator.

Default: use global random seed

Parameter for:

-sfcknn.seed <long>

Random generator.

Default: use global random seed

Parameter for:

-sfcknn.variants <int>

Number of curve variants to generate.

Default: 1

Parameter for:

-sfcknn.variants <int>

Number of curve variants to generate.

Default: 1

Parameter for:

-sfcknn.windowmult <double>

Window size multiplicator.

Default: 10.0

Parameter for:

-sfcknn.windowmult <double>

Window size multiplicator.

Default: 10.0

Parameter for:

-sharedNearestNeighbors <int>

number of nearest neighbors to consider (at least 1)

Parameter for:

-shuffle.seed <long>

Seed for randomly shuffling the rows for the database. If the parameter is not set, a random seed will be used.

Default: use global random seed

Parameter for:

-signitrend.bias <double>

Adjustment for chance: a small constant corresponding to background noise levels.

Parameter for:

-signitrend.halflife <double>

Half-life time: number of time steps until data has lost half its weight.

Parameter for:

-signitrend.minsigma <double>

Significance threshold for reporting

Parameter for:

-silhouette.clustering <class|object>

Clustering algorithm to use for the silhouette coefficients.

Class Restriction: implements algorithm.clustering.ClusteringAlgorithm

Known implementations:

Parameter for:

-silhouette.distance <class|object>

Distance function to use for computing the silhouette.

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-silhouette.noisehandling <MERGE_NOISE | TREAT_NOISE_AS_SINGLETONS | IGNORE_NOISE>

Control how noise should be treated.

Default: TREAT_NOISE_AS_SINGLETONS

Parameter for:

-similarityfunction.preprocessor <class|object>

Preprocessor to use.

Class Restriction: implements index.preprocessed.snn.SharedNearestNeighborIndex

Default: SharedNearestNeighborPreprocessor

Known implementations:

Parameter for:

-simmatrix.scaling <class|object>

Class to use as scaling function.

Class Restriction: implements utilities.scaling.ScalingFunction

Known implementations:

Parameter for:

-simmatrix.skipzero <|true|false>

Skip zero values when computing the colors to increase contrast.

Default: false

Parameter for:

-sne.perplexity <double>

Desired perplexity (approximately the number of neighbors to preserve)

Default: 40.0

Parameter for:

-sne.sigma <double>

Gaussian kernel standard deviation.

Parameter for:

-snn.epsilon <int>

The minimum SNN density.

Parameter for:

-snn.minpts <int>

Threshold for minimum number of points in the epsilon-SNN-neighborhood of a point.

Parameter for:

-SNNDistanceFunction <class|object>

the distance function to asses the nearest neighbors

Class Restriction: implements distance.distancefunction.DistanceFunction

Default: minkowski.EuclideanDistanceFunction

Known implementations:

Parameter for:

-sod.alpha <double>

The multiplier for the discriminance value for discerning small from large variances.

Default: 1.1

Parameter for:

-sod.knn <int>

The number of most snn-similar objects to use as reference set for learning the subspace properties.

Parameter for:

-sod.models <|true|false>

Report the models computed by SOD (default: report only scores).

Default: false

Parameter for:

-sod.similarity <class|object>

The similarity function used for the neighborhood set.

Class Restriction: implements distance.similarityfunction.SimilarityFunction

Default: SharedNearestNeighborSimilarityFunction

Known implementations:

Parameter for:

-sos.k <int>

Number of neighbors to use. Should be about 3x the desired perplexity.

Default: 15

Parameter for:

-sos.perplexity <double>

Perplexity to use.

Default: 4.5

Parameter for:

-spatial.bulkstrategy <class|object>

The class to perform the bulk split with.

Class Restriction: implements index.tree.spatial.rstarvariants.strategies.bulk.BulkSplit

Known implementations:

Parameter for:

-split.dims <int_1,...,int_n>

Dimensions to split into the first relation.

Parameter for:

-sqrtscale.max <double>

Fixed maximum to use in sqrt scaling.

Parameter for:

-sqrtscale.min <double>

Fixed minimum to use in sqrt scaling.

Parameter for:

-sqrtstddevscale.lambda <double>

Significance level to use for error function.

Default: 3.0

Parameter for:

-sqrtstddevscale.mean <double>

Fixed mean to use in standard deviation scaling.

Parameter for:

-sqrtstddevscale.min <double>

Fixed minimum to use in sqrt scaling.

Parameter for:

-ssq.distance <class|object>

Distance function to use for computing the SSQ.

Class Restriction: implements distance.distancefunction.NumberVectorDistanceFunction

Default: minkowski.SquaredEuclideanDistanceFunction

Known implementations:

Parameter for:

-ssq.noisehandling <MERGE_NOISE | TREAT_NOISE_AS_SINGLETONS | IGNORE_NOISE>

Control how noise should be treated.

Default: TREAT_NOISE_AS_SINGLETONS

Parameter for:

-star.nocenter <|true|false>

Do not use the center as extra reference point.

Default: false

Parameter for:

-star.scale <double>

Scale the reference points by the given factor. This can be used to obtain reference points outside the used data space.

Default: 1.0

Parameter for:

-stddevout.k <int>

Number of neighbors to get for stddev based outlier detection.

Parameter for:

-stddevscale.lambda <double>

Significance level to use for error function.

Default: 3.0

Parameter for:

-stddevscale.mean <double>

Fixed mean to use in standard deviation scaling.

Parameter for:

-string.comment <pattern>

Ignore lines in the input file that satisfy this pattern.

Default: ^\s*(#|//|;).*$

Parameter for:

-string.trim <|true|false>

Remove leading and trailing whitespace from each line.

Default: false

Parameter for:

-subclu.distancefunction <class|object>

Distance function to determine the distance between database objects.

Class Restriction: implements distance.distancefunction.subspace.DimensionSelectingSubspaceDistanceFunction

Default: SubspaceEuclideanDistanceFunction

Known implementations:

Parameter for:

-subclu.epsilon <double>

The maximum radius of the neighborhood to be considered.

Parameter for:

-subclu.mindim <int>

Minimum dimensionality to generate clusters for.

Parameter for:

-subclu.minpts <int>

Threshold for minimum number of points in the epsilon-neighborhood of a point.

Parameter for:

-svm.kernel <LINEAR | QUADRATIC | CUBIC | RBF | SIGMOID>

Kernel to use with SVM.

Default: RBF

Parameter for:

-svm.nu <double>

SVM nu parameter.

Default: 0.05

Parameter for:

-tf.normalize <|true|false>

Normalize vectors to manhattan length 1 (convert term counts to term frequencies)

Default: false

Parameter for:

-thresholdclust.scaling <class|object>

Class to use as scaling function.

Class Restriction: implements utilities.scaling.ScalingFunction

Default: IdentityScaling

Known implementations:

Parameter for:

-thresholdclust.threshold <double_1,...,double_n>

Threshold(s) to apply.

Parameter for:

-time <|true|false>

Enable logging of runtime data. Do not combine with more verbose logging, since verbose logging can significantly impact performance.

Default: false

Parameter for:

-tma.p <double>

the percentile parameter

Parameter for:

-tooltip.digits <int>

Number of digits to show (e.g. when visualizing outlier scores)

Default: 4

Parameter for:

-topk.binary <|true|false>

Make the top k a binary scaling.

Default: false

Parameter for:

-topk.k <int>

Number of outliers to keep.

Parameter for:

-trimmedestimate.inner <class|object>

Estimator to use on the trimmed data.

Class Restriction: implements math.statistics.distribution.estimator.DistributionEstimator

Known implementations:

Parameter for:

-trimmedestimate.trim <double>

Relative amount of data to trim on each end, must be 0 < trim < 0.5

Parameter for:

-tsne.affinity <class|object>

Affinity matrix builder.

Class Restriction: implements algorithm.projection.AffinityMatrixBuilder

Default: NearestNeighborAffinityMatrixBuilder

Known implementations:

Parameter for:

-tsne.affinity <class|object>

Affinity matrix builder.

Class Restriction: implements algorithm.projection.AffinityMatrixBuilder

Default: PerplexityAffinityMatrixBuilder

Known implementations:

Parameter for:

-tsne.dim <int>

Output dimensionality.

Default: 2

Parameter for:

-tsne.dim <int>

Output dimensionality.

Default: 2

Parameter for:

-tsne.iter <int>

Number of iterations to perform.

Default: 1000

Parameter for:

-tsne.iter <int>

Number of iterations to perform.

Default: 1000

Parameter for:

-tsne.learningrate <double>

Learning rate of the method.

Default: 200.0

Parameter for:

-tsne.learningrate <double>

Learning rate of the method.

Default: 200.0

Parameter for:

-tsne.momentum <double>

The final momentum to use.

Default: 0.8

Parameter for:

-tsne.momentum <double>

The final momentum to use.

Default: 0.8

Parameter for:

-tsne.retain-original <|true|false>

Retain the original data.

Default: false

Parameter for:

-tsne.seed <long>

Random generator seed

Default: use global random seed

Parameter for:

-tsne.seed <long>

Random generator seed

Default: use global random seed

Parameter for:

-tsne.theta <double>

Approximation quality parameter

Default: 0.5

Parameter for:

-uncertain.dimensionality <int>

Dimensionality of the data set (used for splitting).

Parameter for:

-uncertain.dimensionality <int>

Dimensionality of the data set (used for splitting).

Parameter for:

-uncertain.probability.column <int>

Column in which the probability is stored, starting at 0. -1 is the last column.

Parameter for:

-uo.discrete.generator <class|object>

Class to generate the point distribution.

Class Restriction: implements data.uncertain.uncertainifier.Uncertainifier

Known implementations:

Parameter for:

-uo.quantity.max <int>

Maximum points per uncertain object.

Default: 10

Parameter for:

-uo.quantity.min <int>

Minimum points per uncertain object (defaults to maximum.

Parameter for:

-uo.symmetric <|true|false>

Generate a symetric uncertain region, centered around the exact data.

Default: false

Parameter for:

-uo.uncertainty.max <double>

Maximum deviation of uncertain bounding box.

Parameter for:

-uo.uncertainty.max3sigma <double>

Maximum 3-sigma deviation of uncertain region.

Parameter for:

-uo.uncertainty.min <double>

Minimum deviation of uncertain bounding box.

Default: 0.0

Parameter for:

-uo.uncertainty.min3sigma <double>

Minimum 3-sigma deviation of uncertain region.

Default: 0.0

Parameter for:

-uofilter.generator <class|object>

Generator to derive uncertain objects from certain vectors.

Class Restriction: implements data.uncertain.uncertainifier.Uncertainifier

Known implementations:

Parameter for:

-uofilter.keep <|true|false>

Keep the original data as well.

Default: false

Parameter for:

-uofilter.seed <long>

Random seed for uncertainification.

Default: use global random seed

Parameter for:

-vafile.partitions <int>

Number of partitions to use in each dimension.

Parameter for:

-vafile.partitions <int>

Number of partitions to use in each dimension.

Parameter for:

-validateknn.force-linear <|true|false>

Force the use of linear scanning as reference.

Default: false

Parameter for:

-validateknn.k <int>

Number of neighbors to retreive for kNN benchmarking.

Parameter for:

-validateknn.pattern <pattern>

Pattern to select query points.

Parameter for:

-validateknn.query <class|object>

Data source for the queries. If not set, the queries are taken from the database.

Class Restriction: implements datasource.DatabaseConnection

Known implementations:

Parameter for:

-validateknn.random <long>

Random generator for sampling.

Default: use global random seed

Parameter for:

-validateknn.sampling <double>

Sampling size parameter. If the value is less or equal 1, it is assumed to be the relative share. Larger values will be interpreted as integer sizes. By default, all data will be used.

Parameter for:

-verbose <|true|false>

Enable verbose messages.

Default: false

Parameter for:

-vis.enable <pattern>

Visualizers to enable by default.

Parameter for:

-vis.format <SVG | PNG | PDF | PS | EPS | JPEG>

File format. Note that some formats requrie additional libraries, only SVG and PNG are default.

Default: SVG

Parameter for:

-vis.maxdim <int>

Maximum number of dimensions to display.

Default: 10

Parameter for:

-vis.output <file>

The output folder.

Parameter for:

-vis.ratio <double>

The width/heigh ratio of the output.

Default: 1.33

Parameter for:

-vis.sampling <int>

Maximum number of objects to visualize by default (for performance reasons).

Default: 10000

Parameter for:

-vis.window.single <|true|false>

Embed visualizers in a single window, not using thumbnails and detail views.

Default: false

Parameter for:

-vis.window.title <string>

Title to use for visualization window.

Parameter for:

-visualizer.stylesheet <string>

Style properties file to use, included properties: classic, default, greyscale, neon, presentation, print

Default: default

Parameter for:

-voronoi.mode <VORONOI | DELAUNAY | V_AND_D>

Mode for drawing the voronoi cells (and/or delaunay triangulation)

Default: VORONOI

Parameter for:

-vov.k <int>

The number of nearest neighbors (not including the query point) of an object to be considered for computing its VOV score.

Parameter for:

-vrc.noisehandling <MERGE_NOISE | TREAT_NOISE_AS_SINGLETONS | IGNORE_NOISE>

Control how noise should be treated.

Default: TREAT_NOISE_AS_SINGLETONS

Parameter for:

-winsorize.inner <class|object>

Estimator to use on the winsorized data.

Class Restriction: implements math.statistics.distribution.estimator.DistributionEstimator

Known implementations:

Parameter for:

-winsorize.winsorize <double>

Relative amount of data to winsorize on each end, must be 0 < winsorize < 0.5

Parameter for:

-xmeans.k_min <int>

The minimum number of clusters to find.

Default: 2

Parameter for:

-xmeans.kmeans <class|object>

kMeans algorithm to use.

Class Restriction: implements algorithm.clustering.kmeans.KMeans

Default: KMeansLloyd

Known implementations:

Parameter for:

-xmeans.quality <class|object>

The quality measure to evaluate splits (e.g. AIC, BIC)

Class Restriction: implements algorithm.clustering.kmeans.quality.KMeansQualityMeasure

Known implementations:

Parameter for:

-xmeans.seed <long>

Random seed for splitting clusters.

Default: use global random seed

Parameter for: