Complete List of Filters

Filter Javadoc / Source Code

Interface for filters. A filter is a matcher that does not add new correspondences to the alignment but instead further processes the given alignment by (1) removing correspondences and/or (2) adding new feature weights to existing correspondences.

Keywords: Filter

ScaleConfidence Javadoc / Source Code

Scales the correspondence confidence values linearly to an given interval (by default [0,1]).

Keywords: Scale Confidence

ConfidenceCombiner Javadoc / Source Code

Combines the additional confidences and set the overall correspondence confidence to be the mean of the selected confidences.

Keywords: Confidence Combiner

ScaleAdditionalConfidence Javadoc / Source Code

Scales the additional correspondence confidence values (that were produced by other filters/matchers) linearly to a given interval (by default [0,1]). Each additional confidence is scaled separately and only the specified additional confidences are scaled. If all of them should be scaled, then leave the set of keys empty.

Keywords: Scale Additional Confidence

CardinalityFilter Javadoc / Source Code

This filter returns only the alignments with the highest confidence if there are n-to-m matched elements. This might not be the best solution.

Keywords: Cardinality Filter

BaseFilterWithSetComparison Javadoc / Source Code

Basic filter for instances which compares sets like neighbours or properties.

Keywords: Base Filter With Set Comparison

BadHostsFilter Javadoc / Source Code

This filter removes correspondences where the source or target has not the same host of the OntModels. E.g. it removes rdf:type=rdf:type or foaf:knows=foaf:knows

Keywords: Bad Hosts Filter

AnonymousNodeFilter Javadoc / Source Code

This filter removes correspondences where the source or target is a blank node.

Keywords: Anonymous Node Filter

TypeFilter Javadoc / Source Code

Filters only class, instance or property matches.

Keywords: Type Filter

MixedTypFilter Javadoc / Source Code

Asserts a homogenous alignment (i.e. only the same type is matched). For instance, correspondences between instances and classes will be deleted.

Not Allowed (examples):

  • class, instance
  • datatype property, object property
  • rdf property, datatype property

Allowed are only exact matches.

Keywords: Mixed Typ Filter

ReflexiveCorrespondenceFilter Javadoc / Source Code

Removes all reflexive edges (which maps A to A) from an alignment.

Keywords: Reflexive Correspondence Filter

NtoMCorrespondenceFilter Javadoc / Source Code

A filter which removes correspondences where source or target is matched to more than one entity. All such correspondences will be removed. As an example: if alignment looks like

  • A, B
  • C, D
  • C, E
  • F, D

then the last three are removed because C and D are matched multiple times.

Keywords: NtoM Correspondence Filter

ConfidenceCurvatureFilter Javadoc / Source Code

Filters the alignment by computing the inflection point of the sorted confidences. To make it more stable a smoothing (spline interpolation) can be used. Furthermore also the elbow point of the confidences can be used to filter them.

Keywords: Confidence Curvature Filter

TopXFilter Javadoc / Source Code

This filter keeps only the top X correspondences according to confidence. The filter can be configured to be source-based (keep only the top X correspondences for each source node). The filter can be configured to be target-based (keep only the top X correspondences for each target node). The filter can be configured to be size-based (based on the smaller or larger side of the alignment).

Keywords: TopX Filter

ConfidenceFilter Javadoc / Source Code

This filter returns only alignments with confidence greater or equals than a specific threshold. Default is 0.9.

Thresholds can be set per type.

Keywords: Confidence Filter

AdditionalConfidenceFilter Javadoc / Source Code

It filters based on the additional confidence. The key and threshold should be provided.

Keywords: Additional Confidence Filter

CommonPropertiesFilter Javadoc / Source Code

Filter which deletes instance mappings if they have no matched properties in common.

Keywords: Common Properties Filter

SimilarNeighboursFilter Javadoc / Source Code

Checks for each instance mapping, how many already matched neighbours it has.

Source_Subject ——Source_Property——Source_Object | | subjectCorrespondence objectCorrespondence | | Target_Subject ——Target_Property——Target_Object

Keywords: Similar Neighbours Filter

SimilarTypeFilter Javadoc / Source Code

Checks for each instance mapping, how many already matched types it has in common. For comparing a type hierarchy, choose SimilarHierarchyFilter.

Keywords: Similar Type Filter

BagOfWordsSetSimilarityFilter Javadoc / Source Code

Filters individual/instance mappings by comparing literals. The literals are selected by the corresponding properties (leave empty to select all). The set of tokens created for each individual are compared with the SetSimilarity.

Keywords: Bag Of Words Set Similarity Filter

SimilarHierarchyFilter Javadoc / Source Code

Check if already matched individuals have a similar hierarchy (class hierarchy). For different computation methods see SimilarHierarchyFilterApproach.

Keywords: Similar Hierarchy Filter

MaxWeightBipartiteExtractor Javadoc / Source Code

Faster implementation than HungarianExtractor for generating a one-to-one alignment. The implementation is based on http://www.mpi-inf.mpg.de/~mehlhorn/Optimization/bipartite_weighted.ps (page 13-19).

Keywords: Max Weight Bipartite Extractor

NaiveAscendingExtractor Javadoc / Source Code

Naive ascending extraction as shown in “Analyzing Mapping Extraction Approaches” (C. Meilicke, H. Stuckenschmidt). It iterates over the sorted (ascending) correspondences and and uses the correspondence with the highest confidence. Afterwards removes every other correspondence with the same source or target.

Keywords: Naive Ascending Extractor

NaiveDescendingExtractor Javadoc / Source Code

Naive descending extraction as shown in “Analyzing Mapping Extraction Approaches” (C. Meilicke, H. Stuckenschmidt). It iterates over the sorted (descending) correspondences and and uses the correspondence with the highest confidence. Afterwards removes every other correspondence with the same source or target. Previously it was called CardinalityFilter aka GreedyExtractor.

Keywords: Naive Descending Extractor

HungarianExtractor Javadoc / Source Code

This implementation uses the Hungarian algorithm to find a one to one mapping. The runtime highly depends on the lower number of concepts (source or target) of the alignment as well as the number of correspondences. If a better runtime is needed, use MaxWeightBipartiteExtractor.

Keywords: Hungarian Extractor

MachineLearningScikitFilter Javadoc / Source Code

This filter learns and applies a classifier given a training sample and an existing alignment.

Keywords: Machine Learning Scikit Filter

TransformersFilter Javadoc / Source Code

This filter extracts the corresponding text for a resource (with the specified and customizable extractor) given all correspondences in the input alignment. The texts of the two resources are fed into the specified transformer model and the prediction is added in form of a confidence to the correspondence. No filtering is applied in this class.

Keywords: Transformers Filter

TransformersFineTunerHpSearch Javadoc / Source Code

Keywords: Transformers Fine Tuner Hp Search

LLMBase Javadoc / Source Code

This filter asks a LLM which entity of the source fits best to an entity of the target. Correspondences needs to be provided such that candidates are available. It will only keep correspondences which are stated to be useful. The difference to

#LLMBinaryFilter

is that all possible matches arte given to the LLM model.

Keywords: LLM Base

TransformersFineTuner Javadoc / Source Code

This class is used to fine-tune a transformer model based on a generated dataset. In every call to the match method, the training data will be generated and appended to a temporary file. When you call the TransformersFineTuner#finetuneModel() method, then a model is fine-tuned and the training file is deleted.

Keywords: Transformers Fine Tuner

LLMChooseGivenEntityFilter Javadoc / Source Code

This filter asks the LLM given a source entity which is the best target entity (out of the ones in the alignment). Afterwards the same is done in the reversed direction.

Keywords: LLM Choose Given Entity Filter

LLMBinaryFilter Javadoc / Source Code

This filter asks a LLM if a given correspondence is correct or not. It has no information about the other correspondences and each correspondence becomes a prediction example for the LLM. It will add the corresponding confidence to the correspondence such that a filtering afterwards is possible.

Keywords: LLM Binary Filter

RelationTypeFineTuner Javadoc / Source Code

This matcher predicts the relation type given a transformer model. This component do not create new correspondences but refine the relation of given class correspondences.

Keywords: Relation Type Fine Tuner

RelationTypePredictor Javadoc / Source Code

This matcher predicts the relation type given a transformer model. This component do not create new correspondences but refine the relation of given class correspondences.

Keywords: Relation Type Predictor

AlcomoFilter Javadoc / Source Code

Filter which makes and alignment coherent. When using this component, please cite: Christian Meilicke. Alignment Incoherence in Ontology Matching. University Mannheim 2011.

This filter wraps the original implementation.

Keywords: Alcomo Filter

LogMapRepairFilter Javadoc / Source Code

This is the logmap repair filter. More information can be found at the wiki and in the github repository.

In case you want to use this filter, make the dependency of matching-owlapi-matchers to exclude the following:


 <exclusions>
    <exclusion>
        <groupId>com.github.ansell.pellet</groupId>
        <artifactId>pellet-modularity</artifactId>
    </exclusion>
</exclusions>

and also include OWLAPI with version 4.1.3


 <dependency>
    <groupId>net.sourceforge.owlapi</groupId>
    <artifactId>owlapi-distribution</artifactId>
    <version>4.1.3</version>
</dependency>

Keywords: Log Map Repair Filter

automatically generated on 2024-09-29 09:12