The Semantic Time Series Ontology
language en

The Semantic Time Series Ontology

Release: December 13th, 2024

Modified on: Aug 1st, 2025
This version:
https://w3id.org/semts/ontology/110#
Previous version:
https://w3id.org/semts/ontology/101#
Revision:
1.1.0
Authors:
Alexander Graß
Contributors:
Rohit A. Deshmukh
Publisher:
Alexander Graß
Website / Repository:
https://semts-ontology.github.io/SemTS
Download serialization:
JSON-LD RDF/XML N-Triples TTL
License:
https://creativecommons.org/licenses/by/4.0/
Visualization:
Visualize with WebVowl
Evaluation:
Evaluate with OOPS!
Cite as:
Alexander Graß, Rohit A. Deshmukh. The Semantic Time Series Ontology. https://semts-ontology.github.io/SemTS/
Provenance of this page
Ontology Specification Draft

Abstract

SemTS is an ontology to semantically structure insights gained from multivariate time series analyses combined with domain-specific information. The concept of SemTS constitutes a specification of informative data points or intervals within time series data, further referred to as segments. Any segment comprises characteristic knowledge associated with the covered time interval. Examples of such knowledge range from common time series features, and structural particularities such as anomalies or motifs, to apriori information provided by domain experts. A classification and semantic representation of this knowledge enables organized reusability and effective propagation.

Semantic Time Series: Overview back to ToC

This ontology has the following classes and properties.

Classes

Object Properties

Data Properties

Semantic Time Series: Description back to ToC

SemTS is an ontology designed to identify and describe segments within time series data, which are specific data points or intervals that can overlap. These segments encompass characteristic knowledge about the time interval they cover, including common time series features, structural anomalies, motifs, or information provided by domain experts. By classifying and semantically representing this knowledge, SemTS promotes organized reusability and efficient propagation, potentially reducing resource expenditure while enhancing future analyses. It employs established semantic approaches. Examples are DCAT to reference associated time series data, OWL-Time to define the index structure of time series data and segments or ML-Schema to expand the expressiveness regarding data analysis task information. SemTS's design involves categorizing time series knowledge and mapping it to specific intervals and dimensions of time series data. It introduces a class called TimeSeriesSegment to model these segments, extending the DCAT Dataset class to enable segments to be part of other segments. This structure allows for the association of knowledge, such as anomalies, with particular intervals or data points. TimeIndex specifications extend OWL-Time classes, while dimensional details are represented by DataDimension. The segment-wise consideration of knowledge indirectly serves as an index structure, linking meaningful time series data with categorized knowledge. At the highest level of abstraction, time series knowledge is divided into three categories: DataKnowledge, ScenarioKnowledge, and MethodKnowledge. DataKnowledge refers to insights extracted directly from the data or through analytical methods, such as class membership from time series clustering. ScenarioKnowledge describes verified contexts, including data annotations or domain-specific process knowledge, often equating to expert-provided a priori information and can also define facts derived from inferred knowledge. MethodKnowledge encompasses effective analytical method presets or mathematical/logical equivalents of established process information.

Examples back to ToC

Cross-reference for Semantic Time Series classes, object properties and data properties back to ToC

This section provides details for each class and property defined by Semantic Time Series.

Classes

Algorithm Recommendationc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#Algorithm

Refers to any algorithmic solution in the context of time series analysis. While this class extends the mls:Algorithm class, it enables a reference to further details on specific task specifics and experiments. It is considered method knowledge as this class can be used to define knowledge about an appropriate or optimal algorithmic solution associated with a time series segment.
has super-classes
Algorithm c, Method Knowledge c
is disjoint with
Parameter Recommendation c

Artificial Time Series Segmentc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#ArtificialTimeSeriesSegment

A time series segment that does not, or only partially refer to concrete time series data. It represents an artificially generated time series segment that might for instance describe the result of a forecast with a time index not being in the range of the original time series data.
has super-classes
Time Series Segment c

Data Dimensionc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#DataDimension

Defines a data dimension (or also referred to as feature or attribute) in a time series segment.
is in domain of
description dp, dimension position dp, dimension unit op, title dp, value datatype op
is in range of
segment dimension op

Data Knowledgec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#DataKnowledge

Defines any knowledge that is derived from the data. This includes statistics, other signal characteristics or knowledge generated thorugh data analysis methods. Examples for the latter case are predictions, anomalies or cluster group memberships.
has super-classes
Knowledge c
has sub-classes
Model Instance c, Segment Data Characteristic c, Segment Prediction c, Segment Relation c
is disjoint with
Knowledge Group c, Method Knowledge c, Scenario Knowledge c

Embedded Valuec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#EmbeddedValue

Refers to a value that is directly stored in the knowledge graph. Compared to referenced values, these values are rather single, non-complex values having a primitive type. One example could be the mean value of a univariate time series.
has super-classes
Value c
is in domain of
value string dp

Knowledgec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#Knowledge

Represents the highest level of any knowledge associated with time series data. Knowledge can be made more specific via subclasses or associated concepts.
has super-classes
has sub-classes
Data Knowledge c, Knowledge Group c, Method Knowledge c, Scenario Knowledge c
is in domain of
description dp, has value op, knowledge concept op, knowledge generation entity op, knowledge quality measure op, title dp
is in range of
group knowledge op, input knowledge op, multi segment knowledge op, output knowledge op, segment knowledge op

Knowledge Comparison Measurec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeComparisonMeasure

Any knowledge that is generated by a comparison of compatible inputs. This might either be raw data or already defined knowledge. A simple example might be the result of a Euclidean distance.
has super-classes
Knowledge Generation Entity c
is disjoint with
Knowledge Generation Method c, Knowledge Reasoning c

Knowledge Conceptc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeConcept

An SKOS concept associated with the knowledge class or any corresponding subclass
has super-classes
Concept c
is in range of
knowledge concept op

Knowledge Generation Conceptc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeGenerationConcept

An SKOS concept associated with the knowledge generation entity class or any corresponding subclass
has super-classes
Concept c
is in range of
knowledge generation concept op

Knowledge Generation Entityc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeGenerationEntity

Represents the highest entity to define the process of knowledge generation. A more specific description can be provided by subclasses or assigned concepts. This class inherits from prov:Agent and thus allows for provenance during knowledge propagation.
has super-classes
Activity c
has sub-classes
Knowledge Comparison Measure c, Knowledge Generation Method c, Knowledge Reasoning c
is in domain of
description dp, generated op, knowledge generation concept op, title dp, used op
is in range of
knowledge generation entity op

Knowledge Generation Inputc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeGenerationInput

Defines the input of a knowledge generation entity. While outputs can only contain produced knowledge, inputs might either reference knowledge or raw data.
has super-classes
Knowledge Generation IO c
is in domain of
input data reference op, input knowledge op
is in range of
used op, was derived from op
is disjoint with
Knowledge Generation Output c

Knowledge Generation IOc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeGenerationIO

Defines any input and output of a knowledge generation entity. It inherits from prov:Entity to define instances that are used by or generated by a prov:Activity, which in turn is the parental class of a knowledge generation entity.
has super-classes
Entity c
has sub-classes
Knowledge Generation Input c, Knowledge Generation Output c
is in domain of
description dp, title dp

Knowledge Generation Methodc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeGenerationMethod

Defines a knowledge generation entity, which corresponds to method-based solution. Compared to other knowledge generation entities this class is a subclass of Algorithm and thus can be linked to algorithmic details.
has super-classes
Algorithm c, Knowledge Generation Entity c
is disjoint with
Knowledge Comparison Measure c, Knowledge Reasoning c

Knowledge Generation Outputc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeGenerationOutput

Defines the output of a knowledge generation entity. Compared to inputs, outputs only allow for a reference to generated knowledge and not to raw data. This, on purpose, excludes data transformations from the overall ontology concept by only focusing on a propagation of knowledge.
has super-classes
Knowledge Generation IO c
is in domain of
output knowledge op, was derived from op
is in range of
generated op
is disjoint with
Knowledge Generation Input c

Knowledge Groupc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeGroup

A grouping of knowledge which itself is again considered knowledge. Can be used to structure knowledge and create hierarchies.
has super-classes
Knowledge c
is in domain of
group knowledge op
is disjoint with
Data Knowledge c, Method Knowledge c, Scenario Knowledge c

Knowledge Quality Measurec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeQualityMeasure

Defines the goodness or confidence of time series knowledge associated with one or multiple segments. For knowledge produced by a knowledge generation method the quality measure is often equal to the evaluation metric of the generation method, which is why this class inherits from mls:EvaluationMeasure.
has super-classes
Evaluation Measure c
is in domain of
description dp, has value op, quality measure concept op, title dp
is in range of
knowledge quality measure op

Knowledge Reasoningc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#KnowledgeReasoning

To complement any knowledge that is generated via analytical methods, this class represents knowledge generated from reasoning processes. This also includes validation processes from human experts or results from reasoning tools. While many analysis results can be regarded unvalidated predictions, this class defines subsequent evaluations to further derive concrete facts or confirmations.
has super-classes
Knowledge Generation Entity c
is disjoint with
Knowledge Comparison Measure c, Knowledge Generation Method c

Method Knowledgec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#MethodKnowledge

Any knowledge that defines an appropriate setting, regarding a particular method that is beneficial or even optimal when analyzing a segment given a particular task.
has super-classes
Knowledge c
has sub-classes
Algorithm Recommendation c, Parameter Recommendation c
is disjoint with
Data Knowledge c, Knowledge Group c, Scenario Knowledge c

Model Instancec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#ModelInstance

This class describes a trained model in the traditional machine learning sense. It inherits from mls:Model and serves as a knowledge entity for subsequent predictions.
has super-classes
Data Knowledge c, Model c
is disjoint with
Segment Data Characteristic c, Segment Prediction c, Segment Relation c

Parameter Recommendationc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#Hyperparameter

A hyperparameter is considered method knowledge as this class can be used to define knowledge about an appropriate or optimal hyperparameter setting associated with the analysis of a time series segment for a concrete task.
has super-classes
Hyperparameter c, Method Knowledge c
is disjoint with
Algorithm Recommendation c

Quality Measure Conceptc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#QualityMeasureConcept

An SKOS concept associated with the quality measure class
has super-classes
Concept c
is in range of
quality measure concept op

Referenced Valuec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#ReferencedValue

Refers to a value that is of arbitrary complexity and not directly stored in the knowledge graph. This class inherits from dcat:Dataset and can be retrieved by the associated metainformation. One example could be a set of generated shapelets defining representative segments of a time series.
has super-classes
Dataset c, Value c

Scenario Knowledgec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#ScenarioKnowledge

Defines any knowledge that is associated with a particular scenario or domain. Scenario knowledge corresponds to validated expert knowledge or scenario-specific facts. It includes simple notes, predefined labels or mathematical expressions describing the underlying time series data.
has super-classes
Knowledge c
is disjoint with
Data Knowledge c, Knowledge Group c, Method Knowledge c

Segment Data Characteristicc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#SegmentDataCharacteristic

Defines analytical, data-specific insights including derived statistics or other structural characteristics such as representative kernel expressions. Knowledge of this class mostly equals extracted data features.
has super-classes
Data Knowledge c
is disjoint with
Model Instance c, Segment Prediction c, Segment Relation c

Segment Predictionc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#SegmentPrediction

Defines knowledge derived from data-driven analyses. This for instance includes anomalies and predictions.
has super-classes
Data Knowledge c
is disjoint with
Model Instance c, Segment Data Characteristic c, Segment Relation c

Segment Relationc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#SegmentRelation

SegmentRelation defines relations between multiple segment. Such a relation could for instance be a cross-correlation or the result of a distance metric.
has super-classes
Data Knowledge c
is disjoint with
Model Instance c, Segment Data Characteristic c, Segment Prediction c

Time Series Segmentc back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#TimeSeriesSegment

Represents a segment (also referred to as slice) of a uni- or multivariate time series. Any segment can correspond to the whole time series, a subintervall potentially further limited to a subset of dimensions or only to a single data point. This class inherits from dcat:Dataset and is enriched by an additional time index, information about represented data dimensions and information about associated time series knowledge.
has super-classes
Dataset c
has sub-classes
Artificial Time Series Segment c
is in domain of
column index dp, is part of op, multi segment knowledge op, row index dp, segment dimension op, segment index op, segment knowledge op
is in range of
input data reference op, is part of op

Valuec back to ToC or Class ToC

IRI: https://w3id.org/semts/ontology#Value

This abstract class serves as parental class of EmbeddedValue and ReferencedValue.
has sub-classes
Embedded Value c, Referenced Value c
is in domain of
dimension unit op, value datatype op
is in range of
has value op

Object Properties

dimension unitop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#hasUnit

dimensionUnit is the relation, which assigns a qudt:Unit instance to a DataDimension or a subclass of Value
has domain
Data Dimension c or Value c
has range
Unit c

generatedop back to ToC or Object Property ToC

IRI: http://www.w3.org/ns/prov#generated

Explicit redefinition for a reuse in SemTS
has sub-properties
generated op
has domain
Activity c
has range
Entity c

generatedop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#generated

Property reusing prov:generated
has super-properties
generated op
has domain
Knowledge Generation Entity c
has range
Knowledge Generation Output c

group knowledgeop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#groupKnowledge

References any knowledge within a particular knowledge group.
has super-properties
top Object Property op
has domain
Knowledge Group c
has range
Knowledge c

has valueop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#hasValue

Relationship to the Value class and thus to allow for a reference to stored data.

has characteristics: functional

has super-properties
top Object Property op
has domain
Knowledge c or Knowledge Quality Measure c
has range
Value c

input data referenceop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#inputDataReference

Defines the relation from a Knowledge Generation Input to a Time Series Segment and thus to metadata which specifies the underlying data.

has characteristics: functional

has super-properties
top Object Property op
has domain
Knowledge Generation Input c
has range
Time Series Segment c

input knowledgeop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#inputKnowledge

Similar to the Input Data Reference, this relation defines all the knowledge associated with the input of a Knowledge Generation Entity.
has super-properties
top Object Property op
has domain
Knowledge Generation Input c
has range
Knowledge c

is part ofop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#isPartOf

Defines a hierarchic relationship between time series segments. The intuition of this property is to differentiate between complete time series, in a sense that it includes the data from start to end and subsegments which are reduced in time or dimensionality and thus can be considered a slice of this original data.

has characteristics: transitive

has super-properties
top Object Property op
has domain
Time Series Segment c
has range
Time Series Segment c

knowledge conceptop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#knowledgeConcept

Relates a concept for a Knowledge instance. Can be used to assign concrete instances from the SemTS taxonomy.

has characteristics: functional

has super-properties
top Object Property op
has domain
Knowledge c
has range
Knowledge Concept c

knowledge generation conceptop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#knowledgeGenerationConcept

Relates a concept for a Knowledge Generation instance.

has characteristics: functional

has domain
Knowledge Generation Entity c
has range
Knowledge Generation Concept c

knowledge generation entityop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#knowledgeGenerationEntity

Relates the Knowledge Generation Entity, which produced this Knowledge instance.

has characteristics: functional

has super-properties
top Object Property op
has domain
Knowledge c
has range
Knowledge Generation Entity c

knowledge quality measureop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#knowledgeQualityMeasure

Allows to assign a quality measure to instantiated knowledge.
has super-properties
top Object Property op
has domain
Knowledge c
has range
Knowledge Quality Measure c

multi segment knowledgeop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#multiSegmentKnowledge

Relates any shared Knowledge available for a set of instances from the class Time Series Segment. One example is the generation of Knowledge that is not specific to a single, but to a group of time series segments.
has super-properties
top Object Property op
has domain
Time Series Segment c
has range
Knowledge c

output knowledgeop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#outputKnowledge

Opposed to Input Knowledge, this property defines the knowledge included in the output of a Knowledge generation Entity.
has super-properties
top Object Property op
has domain
Knowledge Generation Output c
has range
Knowledge c

quality measure conceptop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#qualityMeasureConcept

Defines the concept of a particular Knowledge Quality Measure.

has characteristics: functional

has super-properties
top Object Property op
has domain
Knowledge Quality Measure c
has range
Quality Measure Concept c

segment dimensionop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#segmentDimension

Defines the dimensions associated with a Time Series Segment. Multiple dimensions indicate that the segment corresponds to a multivariate time series.
has super-properties
top Object Property op
has domain
Time Series Segment c
has range
Data Dimension c

segment indexop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#segmentIndex

Defines the index associated with a Time Series Segment. Inherits from the time:TemporalEntity.

has characteristics: functional

has super-properties
top Object Property op
has domain
Time Series Segment c
has range
Temporal Entity c

segment knowledgeop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#segmentKnowledge

Relates any Knowledge available for a particular Time Series Segment.
has super-properties
top Object Property op
has domain
Time Series Segment c
has range
Knowledge c

usedop back to ToC or Object Property ToC

IRI: http://www.w3.org/ns/prov#used

Explicit redefinition for a reuse in SemTS
has sub-properties
used op
has domain
Activity c
has range
Entity c

usedop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#used

Property reusing prov:used
has super-properties
used op
has domain
Knowledge Generation Entity c
has range
Knowledge Generation Input c

value datatypeop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#hasDatatype

Assigns a qudt:Datatype instance to an instance of a Value subclass or DataDimension
has domain
Data Dimension c or Value c
has range
Datatype c

was derived fromop back to ToC or Object Property ToC

IRI: http://www.w3.org/ns/prov#wasDerivedFrom

Explicit redefinition for a reuse in SemTS
has sub-properties
was derived from op
has domain
Entity c
has range
Entity c

was derived fromop back to ToC or Object Property ToC

IRI: https://w3id.org/semts/ontology#wasDerivedFrom

Property reusing prov:wasDerivedFrom
has super-properties
was derived from op
has domain
Knowledge Generation Output c
has range
Knowledge Generation Input c

Data Properties

column indexdp back to ToC or Data Property ToC

IRI: https://w3id.org/semts/ontology#datasetColumnIndex

Specifies the column index of a time series segment within a data file or database, in case a source contains multiple time series (e.g. nested dataframes)
has domain
Time Series Segment c
has range
integer

descriptiondp back to ToC or Data Property ToC

IRI: https://w3id.org/semts/ontology#description

Can be used to optionally add a description to an entity of the domain.
has super-properties
top Data Property dp
has domain
Data Dimension c or Knowledge c or Knowledge Generation Entity c or Knowledge Generation IO c or Knowledge Quality Measure c
has range
string

dimension positiondp back to ToC or Data Property ToC

IRI: https://w3id.org/semts/ontology#dimensionPosition

Specifies the position of the associated dimension within a dataframe or similar ordered data structures.

has characteristics: functional

has super-properties
top Data Property dp
has domain
Data Dimension c
has range
int

row indexdp back to ToC or Data Property ToC

IRI: https://w3id.org/semts/ontology#datasetRowIndex

Specifies the row index of a time series segment within a data file or database, in case a source contains multiple time series (e.g. nested dataframes)
has domain
Time Series Segment c
has range
integer

titledp back to ToC or Data Property ToC

IRI: https://w3id.org/semts/ontology#title

Can be used to optionally add a title to an entity of the domain.

has characteristics: functional

has super-properties
top Data Property dp
has domain
Data Dimension c or Knowledge c or Knowledge Generation Entity c or Knowledge Generation IO c or Knowledge Quality Measure c
has range
string

value stringdp back to ToC or Data Property ToC

IRI: https://w3id.org/semts/ontology#valueString

The string representation for a concrete value of an Embedded Value. Complementary, the actual data type is defined via the dType property.

has characteristics: functional

has super-properties
top Data Property dp
has domain
Embedded Value c
has range
string

Legend back to ToC

c: Classes
op: Object Properties
dp: Data Properties

Changes from last version

Classes

Modified classes
Deleted classes

Object Properties

Modified object properties
Added object properties
Deleted object properties

Data Properties

Modified data properties
  • https://w3id.org/semts/ontology#description
    • Added: rdfs:comment "Can be used to optionally add a description to an entity of the domain."@en
    • Added: rdfs:seeAlso "dcterms:description"^^xsd:string
    • Deleted: rdfs:comment "Inherits from dc:description and can be used to optionally add a description to an entity of the domain."@en
  • https://w3id.org/semts/ontology#title
    • Added: rdfs:comment "Can be used to optionally add a title to an entity of the domain."@en
    • Added: rdfs:seeAlso "dcterms:title"^^xsd:string
    • Deleted: rdfs:comment "Inherits from dc:title and can be used to optionally add a title to an entity of the domain."@en
Added data properties

Acknowledgments back to ToC

The authors would like to thank Silvio Peroni for developing LODE, a Live OWL Documentation Environment, which is used for representing the Cross Referencing Section of this document and Daniel Garijo for developing Widoco, the program used to create the template used in this documentation.