DL-Learner Components

Click on the following items to filter the listing below by implemented interfaces (requires Javascript):

show all

Click on a component to get an overview on its configuration options.

AMeasure org.dllearner.accuracymethods.AccMethodAMeasure

short name
ameasure
version
0.1
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
betabeta factor (0 = do not use) double0 false

AMeasure Approximate org.dllearner.accuracymethods.AccMethodAMeasureApprox

short name
approx.ameasure
version
0.1
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
approxDeltaThe Approximate Delta double0.05 false
betabeta factor (0 = do not use) double0 false
reasoner(configured by learning problem) Reasoner false

BUNDLE org.dllearner.core.probabilistic.unife.BUNDLE

short name
bundle
version
1.0
implements
  • OtherComponent
option namedescriptiontypedefault valuerequired?
accuracyaccuracy used during the computation of the probabilistic values (number of digital places) int5 false
bddFTypelibrary used for BDD compilation Stringbuddy false
maxExplanationsthe maximum number of explanations to find for each query int2147483647 false
timeoutmax time allowed for the inference (format: [0-9]h[0-9]m[0-9]s) String0s (infinite timeout) false

CELOE org.dllearner.algorithms.celoe.CELOE

short name
celoe
version
1.0
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
description
CELOE is an adapted and extended version of the OCEL algorithm applied for the ontology engineering use case. See http://jens-lehmann.org/files/2011/celoe.pdf for reference.
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
expandAccuracy100Nodeswhether to try and refine solutions which already have accuracy value of 1 booleanfalse false
filterDescriptionsFollowingFromKBIf true, then the results will not contain suggestions, which already follow logically from the knowledge base. Be careful, since this requires a potentially expensive consistency check for candidate solutions. booleanfalse false
heuristicno description available AbstractHeuristicceloe_heuristic false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
maxClassExpressionTestsThe maximum number of candidate hypothesis the algorithm is allowed to test (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.) int0 false
maxClassExpressionTestsAfterImprovementThe maximum number of candidate hypothesis the algorithm is allowed after an improvement in accuracy (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.) int0 false
maxDepthmaximum depth of description double7 false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
maxExecutionTimeInSecondsAfterImprovementmaximum execution of the algorithm in seconds after last improvement int0 false
maxNrOfResultsSets the maximum number of results one is interested in. (Setting this to a lower value may increase performance as the learning algorithm has to store/evaluate/beautify less descriptions). int10 false
noisePercentagethe (approximated) percentage of noise within the examples double0.0 false
operatorthe refinement operator instance to use LengthLimitedRefinementOperator false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
replaceSearchTreespecifies whether to replace the search tree in the log file after each run or append the new search tree booleanfalse false
reuseExistingDescriptionIf true, the algorithm tries to find a good starting point close to an existing definition/super class of the given class in the knowledge base. booleanfalse false
searchTreeFilefile to use for the search tree Stringlog/searchTree.txt false
singleSuggestionModeUse this if you are interested in only one suggestion and your learning problem has many (more than 1000) examples. booleanfalse false
startClassYou can specify a start class for the algorithm. To do this, you have to use Manchester OWL syntax either with full IRIs or prefixed IRIs. Example: ex:Male or http://example.org/ontology/Female OWLClassExpressionowl:Thing false
stopOnFirstDefinitionalgorithm will terminate immediately when a correct definition is found booleanfalse false
terminateOnNoiseReachedspecifies whether to terminate when noise criterion is met booleanfalse false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false
writeSearchTreespecifies whether to write a search tree booleanfalse false

Class as Instance LP org.dllearner.learningproblems.ClassAsInstanceLearningProblem

short name
classasinstance
version
0.1
implements
  • LearningProblem
option namedescriptiontypedefault valuerequired?
exampleLoaderHelperload examples via class expression selector ExampleLoader false
negativeExamplesno description available Set false
percentPerLengthUnitPercent Per Length Unit double0.05 false
positiveExamplesno description available Set false

ClassExpressionLearningProblem org.dllearner.learningproblems.ClassExpressionLearningProblem

short name
celp
version
0.6
implements
  • LearningProblem
option namedescriptiontypedefault valuerequired?
accuracyMethodSpecifies, which method/function to use for computing accuracy. Available measues are "pred_acc" (predictive accuracy), "fmeasure" (F measure), "generalised_fmeasure" (generalised F-Measure according to Fanizzi and d'Amato). AccMethodPRED_ACC false
betaEqbeta index for F-measure in definition learning double1.0 false
betaSCbeta index for F-measure in super class learning double3.0 false
checkConsistencywhether to check for consistency of suggestions (when added to ontology) booleantrue false
classExpressionToDescribeOWL class expression of which an OWL class expression should be learned OWLClassExpression true
equivalenceWhether this is an equivalence problem (or superclass learning problem) booleantrue false
exampleLoaderHelperload examples via class expression selector ExampleLoader false
maxExecutionTimeInSecondsMaximum execution time in seconds int10 false

ClassLearningProblem org.dllearner.learningproblems.ClassLearningProblem

short name
clp
version
0.6
implements
  • LearningProblem
option namedescriptiontypedefault valuerequired?
accuracyMethodSpecifies, which method/function to use for computing accuracy. Available measues are "pred_acc" (predictive accuracy), "fmeasure" (F measure), "generalised_fmeasure" (generalised F-Measure according to Fanizzi and d'Amato). AccMethodPRED_ACC false
betaEqbeta index for F-measure in definition learning double1.0 false
betaSCbeta index for F-measure in super class learning double3.0 false
checkConsistencywhether to check for consistency of suggestions (when added to ontology) booleantrue false
classToDescribeclass of which an OWL class expression should be learned IRI true
equivalenceWhether this is an equivalence problem (or superclass learning problem) booleantrue false
exampleLoaderHelperload examples via class expression selector ExampleLoader false
maxExecutionTimeInSecondsMaximum execution time in seconds int10 false

Command Line Interface org.dllearner.cli.CLI

option namedescriptiontypedefault valuerequired?
logLevelConfigure logger log level from conf file. Available levels: "FATAL", "ERROR", "WARN", "INFO", "DEBUG", "TRACE". Note, to see results, at least "INFO" is required. StringINFO false
nrOfFoldsNumber of folds in Cross-Validation mode int10 false
performCrossValidationRun in Cross-Validation mode booleanfalse false

Disjunctive ELTL org.dllearner.algorithms.el.ELLearningAlgorithmDisjunctive

short name
deltl
version
0.5
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
description
Disjunctive ELTL is an algorithm based on the refinement operator in http://jens-lehmann.org/files/2009/el_ilp.pdf with support for disjunctions.
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
instanceBasedDisjointswhether to do real disjoint tests or check that two named classes do not have common instances boolean false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
minimumTreeScorethe minimum quality a tree must have to proceed double-1 false
noisePercentagethe (approximated) percentage of noise within the examples double0.0 false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
startClassYou can specify a start class for the algorithm. To do this, you have to use Manchester OWL syntax without using prefixes. OWLClassExpressionowl:Thing false
stopOnFirstDefinitionalgorithm will terminate immediately when a correct definition is found booleanfalse false
treeSearchTimeSecondsSpecifies how long the algorithm should search for a partial solution (a tree). double1.0 false
tryFullCoverageIf yes, then the algorithm tries to cover all positive examples. Note that while this improves accuracy on the testing set, it may lead to overfitting. booleanfalse false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false

DisjunctiveHeuristic org.dllearner.algorithms.el.DisjunctiveHeuristic

short name
disjunctive_heuristic
version
0.1
implements
  • Heuristic
This component does not have configuration options.

DummyParameterLearner org.dllearner.algorithms.probabilistic.parameter.unife.edge.DummyParameterLearner

short name
dummypl
version
1.0
implements
  • LearningAlgorithm
option namedescriptiontypedefault valuerequired?
accuracyaccuracy used during the computation of the probabilistic values (number of digital places) int5 false
differenceLLstop difference between log-likelihood of two consecutive EM cycles double0.000000000028 false
fixedProbabilityValue of the fixed probability. All the probabilistic axioms will have the same probability double0.4 false
keepParametersIf true EDGE keeps the old parameter values of all the probabilistic axioms and it does not relearn them booleanfalse false
maxExplanationsthe maximum number of explanations to find for each query int2147483647 false
maxIterationsmaximum number of cycles long9223372036854775807 false
maxNegativeExamplesmax number of negative examples that edge must handle when a class learning problem is given int0 (infinite) false
maxPositiveExamplesmax number of positive examples that edge must handle when a class learning problem is given int0 (infinite) false
outputformatformat of the output file PossibleOutputFormatOWLXML false
probabilizeAllmake probabilistic all the axioms in the starting probabilistic ontology (including non probabilistic ones) booleanfalse false
randomizerandomize the starting probabilities of the probabilistic axioms booleanfalse false
ratioLLstop ratio between log-likelihood of two consecutive EM cycles double0.000000000028 false
seedseed for random generation int0 false
showAllforce the visualization of all results booleanfalse false
timeoutmax time allowed for the inference (format: [0-9]h[0-9]m[0-9]s) String0s (infinite timeout) false

EDGE org.dllearner.algorithms.probabilistic.parameter.unife.edge.EDGE

short name
edge
version
1.0
implements
  • LearningAlgorithm
option namedescriptiontypedefault valuerequired?
accuracyaccuracy used during the computation of the probabilistic values (number of digital places) int5 false
differenceLLstop difference between log-likelihood of two consecutive EM cycles double0.000000000028 false
keepParametersIf true EDGE keeps the old parameter values of all the probabilistic axioms and it does not relearn them booleanfalse false
maxExplanationsthe maximum number of explanations to find for each query int2147483647 false
maxIterationsmaximum number of cycles long9223372036854775807 false
maxNegativeExamplesmax number of negative examples that edge must handle when a class learning problem is given int0 (infinite) false
maxPositiveExamplesmax number of positive examples that edge must handle when a class learning problem is given int0 (infinite) false
outputformatformat of the output file PossibleOutputFormatOWLXML false
probabilizeAllmake probabilistic all the axioms in the starting probabilistic ontology (including non probabilistic ones) booleanfalse false
randomizerandomize the starting probabilities of the probabilistic axioms booleanfalse false
ratioLLstop ratio between log-likelihood of two consecutive EM cycles double0.000000000028 false
seedseed for random generation int0 false
showAllforce the visualization of all results booleanfalse false
timeoutmax time allowed for the inference (format: [0-9]h[0-9]m[0-9]s) String0s (infinite timeout) false

EDGEDistributedDynamic org.dllearner.algorithms.probabilistic.parameter.distributed.unife.edge.EDGEDistributedDynamic

short name
edgedynamic
version
1.0
implements
  • LearningAlgorithm
option namedescriptiontypedefault valuerequired?
chunkDimnumber of example for chunk int1 false
maxSenderThreads max number of concurrent threads which send examples to the slaves int#processors - 1 false

EDGEDistributedSingleStep org.dllearner.algorithms.probabilistic.parameter.distributed.unife.edge.EDGEDistibutedSingleStep

short name
edgesingle
version
1.0
implements
  • LearningAlgorithm
This component does not have configuration options.

ELTL org.dllearner.algorithms.el.ELLearningAlgorithm

short name
eltl
version
0.5
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
description
ELTL is an algorithm based on the refinement operator in http://jens-lehmann.org/files/2009/el_ilp.pdf.
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
classToDescribeclass of which an OWL class expression should be learned IRI false
heuristicThe heuristic variable to use for ELTL ELHeuristicStableHeuristic false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
instanceBasedDisjointsSpecifies whether to use real disjointness checks or instance based ones (no common instances) in the refinement operator. booleantrue false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
maxClassExpressionDepthThe maximum depth for class expressions to test int2 false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
maxNrOfResultsSets the maximum number of results one is interested in int10 false
noisePercentagethe (approximated) percentage of noise within the examples double0.0 false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
replaceSearchTreespecifies whether to replace the search tree in the log file after each run or append the new search tree booleanfalse false
searchTreeFilefile to use for the search tree Stringlog/searchTree.txt false
startClassYou can specify a start class for the algorithm. To do this, you have to use Manchester OWL syntax without using prefixes. OWLClassExpressionowl:Thing false
stopOnFirstDefinitionalgorithm will terminate immediately when a correct definition is found booleanfalse false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false
writeSearchTreespecifies whether to write a search tree booleanfalse false

ETDT org.dllearner.algorithms.decisiontrees.dsttdt.DSTTDTClassifier

short name
etdt
version
1.0
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
description
An Evidence-based Terminological Decision Tree
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
beamvalue for limiting the number of generated concepts int4 false
heuristicinstance of heuristic to use TreeInductionHeuristicsTreeInductionHeuristics false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
nonSpecifityControla flag to decide if further control on the purity measure should be made booleanfalse false
operatorrefinement operator instance to use RefinementOperatorDLTreesRefinementOperator false
puritythresholdPurity threshold for setting a leaf double0.05 false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false

ExampleLoader org.dllearner.learningproblems.ExampleLoader

short name
ExampleLoader
version
0.1
implements
  • OtherComponent
description
Load examples from Class Expression
option namedescriptiontypedefault valuerequired?
negativeExamplesCEclass expression of negative examples OWLClassExpression false
negativeRandomCountrandomly choose only so many negative examples int false
positiveExamplesCEclass expression of positive examples OWLClassExpression false
positiveRandomCountrandomly choose only so many positive examples int false
randomSeedrandom seed for deterministic example choice long false

FMeasure org.dllearner.accuracymethods.AccMethodFMeasure

short name
fmeasure
version
0.0
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
betabeta factor (0 = do not use) double0 false

FMeasure Approximate org.dllearner.accuracymethods.AccMethodFMeasureApprox

short name
approx.fmeasure
version
0.0
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
approxDeltaThe Approximate Delta double0.05 false
betabeta factor (0 = do not use) double0 false
reasonerreasoner component (configured by learning problem) Reasoner false

Flexible Heuristic org.dllearner.algorithms.ocel.FlexibleHeuristic

short name
flexheuristic
version
0.1
implements
  • Heuristic
option namedescriptiontypedefault valuerequired?
nrOfNegativeExamplesthe number of negative examples int false
percentPerLengthUnitscore percent to deduct per expression length double true

FuzzyPosNegLPStandard org.dllearner.learningproblems.FuzzyPosNegLPStandard

short name
fuzzyPosNeg
version
0.2
implements
  • LearningProblem
option namedescriptiontypedefault valuerequired?
accuracyMethodSpecifies, which method/function to use for computing accuracy. Available measues are "PRED_ACC" (predictive accuracy), "FMEASURE" (F measure), "GEN_FMEASURE" (generalised F-Measure according to Fanizzi and d'Amato). HeuristicTypePRED_ACC false
fuzzyExamplesno description available SortedSet false
negativeExamplesno description available SortedSet false
positiveExamplesno description available SortedSet false

GLOBAL OPTIONS org.dllearner.cli.DocumentationGeneratorMeta.GlobalDoc

option namedescriptiontypedefault valuerequired?
prefixesMapping of prefixes to replace inside other configuration file entries Example: [ ("ex","http://example.com/father#") ] Map false
renderingThe string renderer for any OWL expression output, can be "dlsyntax" or "manchester" Example: dlsyntax Stringmanchester false

Generalised FMeasure org.dllearner.accuracymethods.AccMethodGenFMeasure

short name
gen_fmeasure
version
0.1
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
betabeta factor (0 = do not use) double0 false

Jaccard Coefficient org.dllearner.accuracymethods.AccMethodJaccard

short name
jaccard
version
0.1
implements
  • AccMethod
This component does not have configuration options.

KB File org.dllearner.kb.KBFile

short name
kbfile
version
0.8
implements
  • KnowledgeSource
option namedescriptiontypedefault valuerequired?
baseDirchange the base directory (must be absolute) Stringdirectory of conf file false
fileNamerelative or absolute path to KB file String false
urlURL pointer to the KB file String false

LEAP org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAP

short name
leap
version
1.0
implements
  • LearningAlgorithm
option namedescriptiontypedefault valuerequired?
accuracyaccuracy used during the computation of the probabilistic values (number of digital places) int5 false
blockSizeGreedySearchthe number of probabilistic axioms that LEAP tries to add into the ontology at each iteration of the greedy search int1 false
classAxiomTypeThis is used to set the type of class axiom to learn. Accepted values (case insensitive): 'subClassOf', 'equivalentClasses', 'both' StringsubClassOf false
dummyClassYou can specify a start class for the algorithm. To do this, you have to use Manchester OWL syntax without using prefixes. IRIowl:learnedClass false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds int10 false

LEAPDistributed org.dllearner.algorithms.probabilistic.structure.distributed.unife.leap.LEAPDistributed

short name
leapdistr
version
1.0
implements
  • LearningAlgorithm
option namedescriptiontypedefault valuerequired?
accuracyaccuracy used during the computation of the probabilistic values (number of digital places) int5 false
differenceLLstop difference between log-likelihood of two consecutive iterations BigDecimal0.00001 false
dummyClassYou can specify a start class for the algorithm. To do this, you have to use Manchester OWL syntax without using prefixes. IRIowl:learnedClass false
maxIterationsmaximum number of iterations long2147000000 false
procPLAnumber of mpi processes of parameter learning algorithm for each probabilistic structure learner process int1 false
procPSLAnumber of mpi processes of probabilistic structure learning algorithm int1 false
ratioLLstop ratio between log-likelihood of two consecutive iterations BigDecimal0.00001 false
targetAxiomsFilenameprobabilistic target axioms which can be deleted from the ontology String false

Lexicographic Heuristic org.dllearner.algorithms.ocel.LexicographicHeuristic

short name
lexheuristic
version
0.1
implements
  • Heuristic
This component does not have configuration options.

Local Endpoint org.dllearner.kb.LocalModelBasedSparqlEndpointKS

short name
local_sparql
version
0.9
implements
  • KnowledgeSource
option namedescriptiontypedefault valuerequired?
cacheDirThe base directory of the SPARQL query cache. Stringtmp folder of the system false
cacheTTLThe time to live in milliseconds for cached SPARQL queries, if enabled. The default value is 86400s(=1 day). long86400 false
defaultGraphURIsa list of default graph URIs List{} false
namedGraphURIsa list of named graph URIs List{} false
pageSizepage size Example: 10000 long10 000 false
queryDelayUse this setting to avoid overloading the endpoint with a sudden burst of queries. A value below 0 means no delay. long50 false
retryCountThe maximum number of retries for the execution of a particular SPARQL query. int3 false
urlURL of the SPARQL endpoint URL true
useCacheUse this setting to enable caching of SPARQL queries in a local database. booleantrue false

Mammalian Phenotype SemKernel Workflow org.dllearner.utilities.semkernel.MPSemKernelWorkflow

short name
mpskw
version
0.1
implements
  • OtherComponent
This component does not have configuration options.

Naive AL Learner org.dllearner.algorithms.NaiveALLearner

short name
naiveALLearner
version
0.1
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
maxLengthmaximum length of class expression int4 false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false

OEHeuristicRuntime org.dllearner.algorithms.celoe.OEHeuristicRuntime

short name
celoe_heuristic
version
0.5
implements
  • Heuristic
option namedescriptiontypedefault valuerequired?
expansionPenaltyFactorpenalty for long descriptions (horizontal expansion) (strong by default) double0.1 false
gainBonusFactorbonus for being better than parent node double0.3 false
nodeRefinementPenaltypenalty if a node description has very many refinements since exploring such a node is computationally very expensive double0.0001 false
startNodeBonusno description available double0.1 false

OWL API Reasoner org.dllearner.reasoning.OWLAPIReasoner

short name
oar
version
0.8
implements
  • ReasonerComponent
option namedescriptiontypedefault valuerequired?
owlLinkURLspecifies the URL of the remote OWLLink server Stringnull false
precomputeClassHierarchyif class hierarchy should be precomputed booleantrue false
precomputeDataPropertyHierarchyno description available booleantrue false
precomputeObjectPropertyHierarchyno description available booleantrue false
precomputeObjectPropertyRangesif object property ranges should be precomputed booleantrue false
precomputePropertyDomainsif property domains should be precomputed booleantrue false
reasonerImplementationspecifies the used OWL API reasoner implementation ReasonerImplementationpellet false
sourcesthe underlying knowledge sources Set true
useFallbackReasonerspecifies whether to use a fallback reasoner if a reasoner call fails because it's not supported or results in a bug. (the fallback works only on the assertional level booleanfalse false
useInstanceCheckswhether to use single instance checks booleanfalse false

OWL Class Expression Learner org.dllearner.algorithms.ocel.OCEL

short name
ocel
version
1.2
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
candidatePostReductionSizemaximum number of candidates to retain int30 false
computeBenchmarkInformationspecifies whether to compute and log benchmark information booleanfalse false
expansionPenaltyFactorFor the MultiHeuristic: how much accuracy gain is worth an increase of horizontal expansion by one (typical value: 0.01) double0.02 false
forceRefinementLengthIncreaseif this variable is set to true, then the refinement operator is applied until all concept of equal length have been found e.g. TOP -> A1 -> A2 -> A3 is found in one loop; the disadvantage are potentially more method calls, but the advantage is that the algorithm is better in locating relevant concept in the subsumption hierarchy (otherwise, if the most general concept is not promising, it may never get expanded) booleantrue false
guaranteeXgoodDescriptionshow many sufficient solutions must be found before termination, if terminateOnNoiseReached is enabled int1 false
heuristicthe heuristic to guide the search ExampleBasedHeuristicMultiHeuristic false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
improveSubsumptionHierarchyif enabled, modifies the subsumption hierarchy such that for each class, there is only a single path to reach it via upward and downward refinement respectively. booleantrue false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
lengthMetricadjust the weights of class expression length in refinement OWLClassExpressionLengthMetricOCEL default metric false
maxClassDescriptionTestsThe maximum number of candidate hypothesis the algorithm is allowed to test (0 = no limit). The algorithm will stop afterwards int0 false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
minExecutionTimeInSecondsMinimum time the algorithm has to run before termination (even if solution already found int0 false
negationPenalty(for the ExampleBasedNode.) penalty value to deduce for using a negated class expression (complementOf) int0 false
negativeWeight(for the ExampleBasedNode.) weighting factor on the number of true negatives (true positives are weigthed with 1) double1.0 false
noisePercentagenoise regulates how many positives can be misclassified and when the algorithm terminates double0.0 false
operatorthe refinement operator instance to use LengthLimitedRefinementOperatorRhoDRDown false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
replaceSearchTreespecifies whether to replace the search tree in the log file after each run or append the new search tree booleanfalse false
searchTreeFilefile to use for the search tree Filelog/searchTree.txt false
showBenchmarkInformationshow additional timing info for benchmark purposes booleanfalse false
startClassYou can specify a start class for the algorithm Example: ex:Male or http://example.org/ontology/Female OWLClassExpressionowl:Thing false
startNodeBonus(for the ExampleBasedNode.) the score value for the start node double0.1 false
terminateOnNoiseReachedspecifies whether to terminate when noise criterion is met booleantrue false
useCandidateReductioncandidate reduction: using this mechanism we can simulate the divide&conquer approach in many ILP programs using a clause by clause search; after a period of time the candidate set is reduced to focus CPU time on the most promising concepts booleantrue false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false
useOverlyGeneralListno description available booleantrue false
usePropernessChecksif set to false we do not test properness; this may seem wrong but the disadvantage of properness testing are additional reasoner queries and a search bias towards ALL r.something because ALL r.TOP is improper and automatically expanded further booleanfalse false
useShortConceptConstructionwhether to shorten concepts to ignore identical refinement. e.g. male AND male is shortened to male. booleantrue false
useTooWeakListexclude too weak concepts when they occur as sub concept booleantrue false
useTreeTraversaltree traversal means to run through the most promising concepts and connect them in an intersection to find a solution (this is called irregularly e.g. every 100 seconds) booleanfalse false
writeSearchTreespecifies whether to write a search tree booleanfalse false

OWL Class Expression Length Metric org.dllearner.utilities.owl.OWLClassExpressionLengthMetric

short name
cel_metric
version
0.1
implements
  • OtherComponent
option namedescriptiontypedefault valuerequired?
classLengthClass: "C" int1 false
dataAllValuesLengthData All Values: "∀" p.t int1 false
dataCardinalityLengthData Cardinality restriction: "≤n" r.t int2 false
dataComplementLengthData Complement: "¬"datatype int1 false
dataHasValueLengthData Has Value: "∃" p."{V}" int2 false
dataIntersectionLengthData Intersection: datatype"⨅"datatype int1 false
dataOneOfLengthData One of: ∃ p."{U,V,W}" int1 false
dataProperyLengthData Property: ∃ "p".t int1 false
dataSomeValuesLengthData Some Values: "∃" p.t int1 false
dataUnionLengthData Union: datatype""datatype int1 false
datatypeLengthDatatype: "^^datatype" int1 false
objectAllValuesLengthObj. All Values: "∀" r.C int1 false
objectCardinalityLengthObj. Cardinality restriction: "≤n" r.C int2 false
objectComplementLengthComplement: "¬"C int1 false
objectHasSelfLengthObj. Self restriction: "∃" r.Self int2 false
objectHasValueLengthObj. Has Value: "∃" r."{I}" int2 false
objectIntersectionLengthIntersection: A"⨅"B int1 false
objectInverseLengthInverse property: ∃ "r⁻".C int2 false
objectOneOfLengthObj. One of: ∃ r."{X,Y,Z}" int1 false
objectProperyLengthObj. Property: ∃ "r".C int1 false
objectSomeValuesLengthObj. Some Values: "∃" r.C int1 false
objectUnionLengthUnion: A"⨆"B int1 false

OWL File org.dllearner.kb.OWLFile

short name
owlfile
version
0.9
implements
  • KnowledgeSource
option namedescriptiontypedefault valuerequired?
baseDirseparately specify directory of KB file String false
defaultGraphURIsa list of default graph URIs to query from the Endpoint List false
fileNamerelative or absolute path to KB file String false
namedGraphURIsa list of named graph URIs to query from the Endpoint List false
reasoningStringEnable JENA reasoning on the Ontology Model. Available reasoners are: "micro_rule", "mini_rule", "rdfs", "rule" Stringfalse false
sparqlSPARQL CONSTRUCT expression to download from Endpoint String false
urlURL pointer to the KB file or Endpoint URL false

OperatorInverter org.dllearner.refinementoperators.OperatorInverter

short name
inv_op
version
0.1
implements
  • RefinementOperator
option namedescriptiontypedefault valuerequired?
guaranteeLengthWhether inverse solutions must respect the desired max length booleantrue false
lengthMetricclass expression length calculation metric OWLClassExpressionLengthMetric false
operatoroperator to invert LengthLimitedRefinementOperator true
useNegationNormalFormwhether to apply NNF booleantrue false

PCELOE org.dllearner.algorithms.celoe.PCELOE

short name
pceloe
version
1.0
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
description
PCELOE is an experimental, parallel implementation of the CELOE algorithm.
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
expandAccuracy100Nodeswhether to try and refine solutions which already have accuracy value of 1 booleanfalse false
filterDescriptionsFollowingFromKBIf true, then the results will not contain suggestions, which already follow logically from the knowledge base. Be careful, since this requires a potentially expensive consistency check for candidate solutions. booleanfalse false
heuristicno description available AbstractHeuristicceloe_heuristic false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
maxClassExpressionTestsThe maximum number of candidate hypothesis the algorithm is allowed to test (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.) int0 false
maxClassExpressionTestsAfterImprovementThe maximum number of candidate hypothesis the algorithm is allowed after an improvement in accuracy (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.) int0 false
maxDepthmaximum depth of description double7 false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
maxExecutionTimeInSecondsAfterImprovementmaximum execution of the algorithm in seconds int0 false
maxNrOfResultsSets the maximum number of results one is interested in. (Setting this to a lower value may increase performance as the learning algorithm has to store/evaluate/beautify less descriptions). int10 false
noisePercentagethe (approximated) percentage of noise within the examples double0.0 false
nrOfThreadsnumber of threads running in parallel int2 false
operatorthe refinement operator instance to use LengthLimitedRefinementOperator false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
replaceSearchTreespecifies whether to replace the search tree in the log file after each run or append the new search tree booleanfalse false
reuseExistingDescriptionIf true, the algorithm tries to find a good starting point close to an existing definition/super class of the given class in the knowledge base. booleanfalse false
searchTreeFilefile to use for the search tree Stringlog/searchTree.txt false
singleSuggestionModeUse this if you are interested in only one suggestion and your learning problem has many (more than 1000) examples. booleanfalse false
startClassYou can specify a start class for the algorithm. To do this, you have to use Manchester OWL syntax without using prefixes. OWLClassExpressionowl:Thing false
stopOnFirstDefinitionalgorithm will terminate immediately when a correct definition is found booleanfalse false
terminateOnNoiseReachedspecifies whether to terminate when noise criterion is met booleanfalse false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false
writeSearchTreespecifies whether to write a search tree booleanfalse false

PosNegLPStandard org.dllearner.learningproblems.PosNegLPStandard

short name
posNegStandard
version
0.8
implements
  • LearningProblem
option namedescriptiontypedefault valuerequired?
accuracyMethodSpecifies, which method/function to use for computing accuracy. Available measues are "PRED_ACC" (predictive accuracy), "FMEASURE" (F measure), "GEN_FMEASURE" (generalised F-Measure according to Fanizzi and d'Amato). AccMethodTwoValuedPRED_ACC false
negativeExampleslist of negative examples Set true
percentPerLengthUnitPercent Per Length Unit double0.05 false
positiveExampleslist of positive examples Set true
useRetrievalForClassification"Specifies whether to use retrieval or instance checks for testing a concept. - NO LONGER FULLY SUPPORTED. booleanfalse false

PosNegLPStrict org.dllearner.learningproblems.PosNegLPStrict

short name
posNegStrict
version
0.1
implements
  • LearningProblem
description
three valued definition learning problem
option namedescriptiontypedefault valuerequired?
accuracyMethodSpecifies, which method/function to use for computing accuracy. Available measues are "PRED_ACC" (predictive accuracy), "FMEASURE" (F measure), "GEN_FMEASURE" (generalised F-Measure according to Fanizzi and d'Amato). AccMethodTwoValuedPRED_ACC false
accuracyPenaltypenalty for pos/neg examples which are classified as neutral double1.0 false
errorPenaltypenalty for pos. examples classified as negative or vice versa double3.0 false
negativeExampleslist of negative examples Set true
penaliseNeutralExamplesif set to true neutral examples are penalised boolean false
percentPerLengthUnitPercent Per Length Unit double0.05 false
positiveExampleslist of positive examples Set true
useRetrievalForClassification"Specifies whether to use retrieval or instance checks for testing a concept. - NO LONGER FULLY SUPPORTED. booleanfalse false

PosNegUndLP org.dllearner.learningproblems.PosNegUndLP

short name
posNegUndLP
version
1.0
implements
  • LearningProblem
description
A learning problem with uncertain-membership instances
option namedescriptiontypedefault valuerequired?
uncertainExamplesthe uncertain examples Set true

Predictive Accuracy org.dllearner.accuracymethods.AccMethodPredAcc

short name
pred_acc
version
0.0
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
betabeta factor (0 = do not use) double0 false

Predictive Accuracy Approximate org.dllearner.accuracymethods.AccMethodPredAccApprox

short name
approx.prec_acc
version
0.0
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
approxDeltaThe Approximate Delta double0.05 false
betabeta factor (0 = do not use) double0 false
reasoner(configured by the learning problem) Reasoner false

Predictive Accuracy without Weak elimination org.dllearner.accuracymethods.AccMethodPredAccOCEL

short name
pred_acc.ocel
version
0.0
implements
  • AccMethod
This component does not have configuration options.

PropertyAxiomLearningProblem org.dllearner.learningproblems.PropertyAxiomLearningProblem

short name
palp
version
0.6
implements
  • LearningProblem
option namedescriptiontypedefault valuerequired?
reasonerThe reasoner component variable to use for this Learning Problem AbstractReasonerComponent false

QueryTreeHeuristic org.dllearner.algorithms.qtl.heuristics.QueryTreeHeuristicSimple

short name
qtree_heuristic_simple
version
0.1
implements
  • Heuristic
This component does not have configuration options.

QueryTreeHeuristicC org.dllearner.algorithms.qtl.heuristics.QueryTreeHeuristicComplex

short name
qtree_heuristic_complex
version
0.1
implements
  • Heuristic
This component does not have configuration options.

Refinement Operator TDT org.dllearner.algorithms.decisiontrees.refinementoperators.DLTreesRefinementOperator

short name
tdtop
version
1.0
implements
  • RefinementOperator
option namedescriptiontypedefault valuerequired?
beamno description available int5 false
lpthe learning problem instance to use PosNegLP false
reasonerthe reasoner instance to use Reasoner false
rono description available int1 false

SPARQL Reasoner org.dllearner.reasoning.SPARQLReasoner

short name
spr
version
0.1
implements
  • ReasonerComponent
option namedescriptiontypedefault valuerequired?
laxModeUse alternative relaxed Sparql-queries for Classes and Individuals booleanfalse false
precomputeClassHierarchyif class hierarchy should be precomputed booleantrue false
precomputeDataPropertyHierarchyno description available booleantrue false
precomputeObjectPropertyHierarchyno description available booleantrue false
precomputeObjectPropertyRangesif object property ranges should be precomputed booleantrue false
precomputePropertyDomainsif property domains should be precomputed booleantrue false
preferAskPrefer ASK queries when there is a choice in implementation booleantrue false
requestLogFileLog file for reasoner request logging String false
requestLoggingLog reasoner requests booleanfalse false
sourcesthe underlying knowledge sources Set true
useGenericSplitsCodeWhether to use the generic facet generation code, which requires downloading all instances and is thus not recommended booleanfalse false
useInstanceCheckswhether to use single instance checks booleanfalse false
useValueListsWhether to use SPARQL1.1 Value Lists booleanfalse false

SPARQL Reasoner (Quad) org.dllearner.reasoning.SPARQLReasonerQuad

short name
spr.quad
version
0.1
implements
  • ReasonerComponent
option namedescriptiontypedefault valuerequired?
laxModeUse alternative relaxed Sparql-queries for Classes and Individuals booleanfalse false
preferAskPrefer ASK queries when there is a choice in implementation booleantrue false
requestLogFileLog file for reasoner request logging String false
requestLoggingLog reasoner requests booleanfalse false
useGenericSplitsCodeWhether to use the generic facet generation code, which requires downloading all instances and is thus not recommended booleanfalse false
useValueListsWhether to use SPARQL1.1 Value Lists booleanfalse false

SPARQL endpoint org.dllearner.kb.SparqlEndpointKS

short name
sparql
version
0.2
implements
  • KnowledgeSource
option namedescriptiontypedefault valuerequired?
cacheDirThe base directory of the SPARQL query cache. Stringtmp folder of the system false
cacheTTLThe time to live in milliseconds for cached SPARQL queries, if enabled. The default value is 86400s(=1 day). long86400 false
defaultGraphURIsa list of default graph URIs List{} false
namedGraphURIsa list of named graph URIs List{} false
pageSizepage size Example: 10000 long10 000 false
queryDelayUse this setting to avoid overloading the endpoint with a sudden burst of queries. A value below 0 means no delay. long50 false
retryCountThe maximum number of retries for the execution of a particular SPARQL query. int3 false
urlURL of the SPARQL endpoint URL true
useCacheUse this setting to enable caching of SPARQL queries in a local database. booleantrue false

SPARQL endpoint fragment org.dllearner.kb.sparql.SparqlKnowledgeSource

short name
sparqlfrag
version
0.5
implements
  • KnowledgeSource
This component does not have configuration options.

SemKernel Workflow org.dllearner.utilities.semkernel.SemKernelWorkflow

short name
skw
version
0.1
implements
  • OtherComponent
This component does not have configuration options.

Stable Heuristic org.dllearner.algorithms.el.StableHeuristic

short name
el_heuristic
version
0.1
implements
  • Heuristic
This component does not have configuration options.

TDT org.dllearner.algorithms.decisiontrees.tdt.TDTClassifier

short name
tdt
version
1.0
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
description
A Terminological Decision Tree
option namedescriptiontypedefault valuerequired?
binaryClassificationif it is a binary classification problem booleanfalse false
ccpvalue for limiting the number of generated concepts booleanfalse false
classToDescribeconcept for splitting undefined examples into positive and negative for binary classification problems OWLClassExpression false
heuristicthe heuristic instance to use TreeInductionHeuristicsTreeInductionHeuristics false
missingValueTreatmentForTDTfor overcoming the problem of missing values in tree algorithms.tree.models booleanfalse false
operatorthe refinement operator instance to use RefinementOperatorDLTreesRefinementOperator false
puritythresholdPurity threshold for setting a leaf double0.05 false

Weighted FMeasure org.dllearner.accuracymethods.AccMethodFMeasureWeighted

short name
weighted.fmeasure
version
0.0
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
balancedbalance the weights to relative set size booleanfalse false
betabeta factor (0 = do not use) double0 false
negWeightweight on the negative examples double1 false
posWeightweight on the positive examples double1 false

Weighted Predictive Accuracy org.dllearner.accuracymethods.AccMethodPredAccWeighted

short name
weighted.pred_acc
version
0.0
implements
  • AccMethod
option namedescriptiontypedefault valuerequired?
balancedbalance the weights to relative set size booleanfalse false
negWeightweight on the negative examples double1 false
posWeightweight on the positive examples double1 false

asymmetric object property axiom learner org.dllearner.algorithms.properties.AsymmetricObjectPropertyAxiomLearner

short name
oplasymm
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for asymmetric object property axioms.
This component does not have configuration options.

closed world reasoner org.dllearner.reasoning.ClosedWorldReasoner

short name
cwr
version
0.9
implements
  • ReasonerComponent
option namedescriptiontypedefault valuerequired?
defaultNegationWhether to use default negation, i.e. an instance not being in a class means that it is in the negation of the class. booleantrue false
forAllSemanticsThis option controls how to interpret the all quantifier in forall r.C. The standard option is to return all those which do not have an r-filler not in C. The domain semantics is to use those which are in the domain of r and do not have an r-filler not in C. The forallExists semantics is to use those which have at least one r-filler and do not have an r-filler not in C. ForallSemanticsstandard false
handlePunningno description available booleanfalse false
materializeExistentialRestrictionsno description available booleanfalse false
precomputeClassHierarchyif class hierarchy should be precomputed booleantrue false
precomputeDataPropertyHierarchyno description available booleantrue false
precomputeObjectPropertyHierarchyno description available booleantrue false
precomputeObjectPropertyRangesif object property ranges should be precomputed booleantrue false
precomputePropertyDomainsif property domains should be precomputed booleantrue false
reasonerComponentthe underlying reasoner implementation OWLAPIReasonerOWL API Reasoner false
sourcesthe underlying knowledge sources Set true
useInstanceCheckswhether to use single instance checks booleanfalse false
useMaterializationCachingno description available booleantrue false

data property domain axiom learner org.dllearner.algorithms.properties.DataPropertyDomainAxiomLearner

short name
dpldomain
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for data property domain axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false

data property range learner org.dllearner.algorithms.properties.DataPropertyRangeAxiomLearner

short name
dblrange
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for reflexive data property range axioms.
This component does not have configuration options.

data subproperty axiom learner org.dllearner.algorithms.properties.SubDataPropertyOfAxiomLearner

short name
dplsubprop
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm data subproperty axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false
betathe beta value for the F-score calculation double1.0 false
strictModeno description available booleanfalse false

disjoint classes learner org.dllearner.algorithms.DisjointClassesLearner

short name
cldisjoint
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
  • ClassExpressionLearningAlgorithm
option namedescriptiontypedefault valuerequired?
entityToDescribethe OWL entity to learn about OWLEntity false
ksthe sparql endpoint knowledge source SparqlEndpointKS false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds (abstract) int10 false
maxFetchedRowsThe maximum number of rows fetched from the endpoint to approximate the result. int false
reasonerThe sparql reasoner instance to use SPARQLReasonerSPARQLReasoner false
returnOnlyNewAxiomsomit axioms already existing in the knowledge base booleanfalse false
suggestMostGeneralClassesonly keep most general classes in suggestions booleantrue false
useClassPopularityinclude instance count / popularity when computing scores booleantrue false

disjoint data properties axiom learner org.dllearner.algorithms.properties.DisjointDataPropertyAxiomLearner

short name
dpldisjoint
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for disjoint data properties axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false
betathe beta value for the F-score calculation double1.0 false
strictModeno description available booleanfalse false

disjoint object properties axiom learner org.dllearner.algorithms.properties.DisjointObjectPropertyAxiomLearner

short name
opldisjoint
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for disjoint object properties axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false
betathe beta value for the F-score calculation double1.0 false
strictModeno description available booleanfalse false

efficient SPARQL fragment extractor org.dllearner.kb.sparql.simple.SparqlSimpleExtractor

short name
sparqls
version
0.1
implements
  • KnowledgeSource
option namedescriptiontypedefault valuerequired?
aboxfilterFilter for the tbox, can use variable ?s, ?p amd ?o String false
defaultGraphURIdefault graph URI String true
endpointURLURL of the SPARQL endpoint String true
instancesList of the instances to use List true
ontologySchemaUrlsList of Ontology Schema URLs List true
recursionDepthrecursion depth int true
sparqlQuerySparql Query String false
tboxfilterFilter for the tbox, can use variable ?example and ?class String false

equivalent data properties axiom learner org.dllearner.algorithms.properties.EquivalentDataPropertyAxiomLearner

short name
dplequiv
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for equivalent data properties axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false
betathe beta value for the F-score calculation double1.0 false
strictModeno description available booleanfalse false

equivalent object properties axiom learner org.dllearner.algorithms.properties.EquivalentObjectPropertyAxiomLearner

short name
oplequiv
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for equivalent object properties axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false
betathe beta value for the F-score calculation double1.0 false
strictModeno description available booleanfalse false

functional data property axiom learner org.dllearner.algorithms.properties.FunctionalDataPropertyAxiomLearner

short name
dplfunc
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for functional data property axioms.
This component does not have configuration options.

functional object property axiom learner org.dllearner.algorithms.properties.FunctionalObjectPropertyAxiomLearner

short name
oplfunc
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for functional object property axioms.
This component does not have configuration options.

inverse functional object property axiom learner org.dllearner.algorithms.properties.InverseFunctionalObjectPropertyAxiomLearner

short name
oplinvfunc
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for inverse functional object property axioms.
This component does not have configuration options.

inverse object property axiom learner org.dllearner.algorithms.properties.InverseObjectPropertyAxiomLearner

short name
oplinv
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for inverse object property axioms.
This component does not have configuration options.

irreflexive object property axiom learner org.dllearner.algorithms.properties.IrreflexiveObjectPropertyAxiomLearner

short name
oplirrefl
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for irreflexive object property axioms.
This component does not have configuration options.

multiple criteria heuristic org.dllearner.algorithms.ocel.MultiHeuristic

short name
multiheuristic
version
0.7
implements
  • Heuristic
option namedescriptiontypedefault valuerequired?
expansionPenaltyFactorhow much accuracy gain is worth an increase of horizontal expansion by one (typical value: 0.01) double0.02 false
gainBonusFactorhow accuracy gain should be weighted versus accuracy itself (typical value: 1.00) double0.5 false
negationPenaltypenalty value to deduce for using a negated class expression (complementOf) int0 false
negativeWeightweighting factor on the number of true negatives (true positives are weigthed with 1) double1.0 false
nodeChildPenaltypenalty factor for the search tree node child count (use higher values for simple learning problems) double0.0001 false
startNodeBonusthe score value for the start node double0.1 false

object property domain axiom learner org.dllearner.algorithms.properties.ObjectPropertyDomainAxiomLearner

short name
opldomain
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for object property domain axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false

object property range learner org.dllearner.algorithms.properties.ObjectPropertyRangeAxiomLearner

short name
oplrange
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for object property range axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false

object subproperty axiom learner org.dllearner.algorithms.properties.SubObjectPropertyOfAxiomLearner

short name
oplsubprop
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm object subproperty axioms.
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false
betathe beta value for the F-score calculation double1.0 false
strictModeno description available booleanfalse false

pattern-based learner org.dllearner.algorithms.pattern.PatternBasedAxiomLearningAlgorithm

short name
patla
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
Pattern-based algorithm uses OWL axioms as pattern.
option namedescriptiontypedefault valuerequired?
entityToDescribethe OWL entity to learn about OWLEntity false
ksthe sparql endpoint knowledge source SparqlEndpointKS false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds (abstract) int10 false
maxFetchedRowsThe maximum number of rows fetched from the endpoint to approximate the result. int false
reasonerThe sparql reasoner instance to use SPARQLReasonerSPARQLReasoner false
returnOnlyNewAxiomsomit axioms already existing in the knowledge base booleanfalse false

positive only learning problem org.dllearner.learningproblems.PosOnlyLP

short name
posonlylp
version
0.6
implements
  • LearningProblem
option namedescriptiontypedefault valuerequired?
exampleLoaderHelperload examples via class expression selector ExampleLoader false
positiveExamplesthe positive examples SortedSet true

query tree learner with noise (disjunctive) org.dllearner.algorithms.qtl.QTL2Disjunctive

short name
qtl2dis
version
0.8
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
betahow important it is not to cover negatives double1 false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
noisePercentagethe (approximated) percentage of noise within the examples double0.0 false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false

query tree learner with noise (disjunctive) - multi-threaded org.dllearner.algorithms.qtl.QTL2DisjunctiveMultiThreaded

short name
qtl2dismt
version
0.8
implements
  • LearningAlgorithm
  • ClassExpressionLearningAlgorithm
option namedescriptiontypedefault valuerequired?
allowedConceptsList of classes that are allowed Set false
allowedDataPropertiesList of data properties to allow Set false
allowedObjectPropertiesList of object properties to allow Set false
betahow important it is not to cover negatives double1 false
ignoredConceptsList of classes to ignore Set false
ignoredDataPropertiesList of data properties to ignore Set false
ignoredObjectPropertiesList of object properties to ignore Set false
learningProblemThe Learning Problem variable to use in this algorithm AbstractClassExpressionLearningProblem false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds long10 false
noisePercentagethe (approximated) percentage of noise within the examples double0.0 false
reasonerThe reasoner variable to use for this learning problem AbstractReasonerComponent false
useMinimizerSpecifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance. booleantrue false

reflexive object property axiom learner org.dllearner.algorithms.properties.ReflexiveObjectPropertyAxiomLearner

short name
oplrefl
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for reflexive object property domain axioms.
This component does not have configuration options.

rho refinement operator org.dllearner.refinementoperators.RhoDRDown

short name
rho
version
0.8
implements
  • RefinementOperator
option namedescriptiontypedefault valuerequired?
applyAllFilterno description available booleantrue false
applyExistsFilterthrowing out all refinements with duplicate ∃ r for any r booleantrue false
cardinalityLimitlimit for cardinality restrictions (this makes sense if we e.g. have compounds with too many atoms) int5 false
disjointChecksskip combination of intersection between disjoint classes booleantrue false
dropDisjunctsif enabled, generalise by removing parts of a disjunction booleanfalse false
frequencyThresholdminimum number an individual or literal has to be seen in the knowledge base before considering it for inclusion in concepts int3 false
instanceBasedDisjointsno description available booleantrue false
lengthMetricclass expression length metric (should match learning algorithm usage) OWLClassExpressionLengthMetricdefault cel_metric false
maxNrOfSplitsthe number of generated split intervals for numeric types int12 false
reasonerthe reasoner to use AbstractReasonerComponent false
startClassYou can specify a start class for the algorithm OWLClassExpressionowl:Thing false
useAllConstructorsupport of universal restrictions (owl:allValuesFrom), e.g. ∀ r.C booleantrue false
useBooleanDatatypessupport of boolean datatypes (xsd:boolean), e.g. ∃ r.{true} booleantrue false
useCardinalityRestrictionssupport of qualified cardinality restrictions (owl:minCardinality, owl:maxCardinality, owl:exactCardinality), e.g. ≥ 3 r.C booleantrue false
useDataHasValueConstructorsupport of has value constructor (owl:hasValue), e.g. ∃ r.{20} booleanfalse false
useExistsConstructorsupport of existential restrictions (owl:someValuesFrom), e.g. ∃ r.C booleantrue false
useHasSelfsupport of local reflexivity of an object property expression (owl:hasSelf), e.g. ∃ loves.Self for a narcissistic booleanfalse false
useHasValueConstructorsupport of has value constructor (owl:hasValue), e.g. ∃ r.{a} booleanfalse false
useInversesupport of inverse object properties (owl:inverseOf), e.g. r⁻.C booleanfalse false
useNegationsupport of negation (owl:complementOf), e.g. ¬ C booleantrue false
useNumericDatatypessupport of numeric datatypes (xsd:int, xsd:double, ...), e.g. ∃ r.{true} booleantrue false
useObjectValueNegationwhether to generate object complement while refining booleanfalse false
useSomeOnlyuniversal restrictions on a property r are only used when there is already a cardinality and/or existential restriction on r booleantrue false
useStringDatatypessupport of string datatypes (xsd:string), e.g. ∃ r.{"SOME_STRING"} booleanfalse false
useTimeDatatypesno description available booleantrue false

simple subclass learner org.dllearner.algorithms.SimpleSubclassLearner

short name
clsub
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
  • ClassExpressionLearningAlgorithm
option namedescriptiontypedefault valuerequired?
batchModecompute everything in a single SPARQL query booleanfalse false
entityToDescribethe OWL entity to learn about OWLEntity false
ksthe sparql endpoint knowledge source SparqlEndpointKS false
maxExecutionTimeInSecondsmaximum execution of the algorithm in seconds (abstract) int10 false
maxFetchedRowsThe maximum number of rows fetched from the endpoint to approximate the result. int false
reasonerThe sparql reasoner instance to use SPARQLReasonerSPARQLReasoner false
returnOnlyNewAxiomsomit axioms already existing in the knowledge base booleanfalse false
strictOWLModeno description available booleanfalse false

symmetric object property axiom learner org.dllearner.algorithms.properties.SymmetricObjectPropertyAxiomLearner

short name
oplsymm
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for symmetric object property axioms.
This component does not have configuration options.

transitive object property axiom learner org.dllearner.algorithms.properties.TransitiveObjectPropertyAxiomLearner

short name
opltrans
version
0.1
implements
  • LearningAlgorithm
  • AxiomLearningAlgorithm
description
A learning algorithm for transitive object property axioms.
This component does not have configuration options.