=Paper=
{{Paper
|id=Vol-2022/paper31
|storemode=property
|title=
A Domain-Agnostic Tool for Scalable Ontology Population and Enrichment from Diverse Linked Data Sources
|pdfUrl=https://ceur-ws.org/Vol-2022/paper31.pdf
|volume=Vol-2022
|authors=Efstratios Kontopoulos,Panagiotis Mitzias,Marina Riga,Ioannis Kompatsiaris
|dblpUrl=https://dblp.org/rec/conf/rcdl/KontopoulosMRK17
}}
==
A Domain-Agnostic Tool for Scalable Ontology Population and Enrichment from Diverse Linked Data Sources
==
https://ceur-ws.org/Vol-2022/paper31.pdf
A Domain-Agnostic Tool for Scalable Ontology
Population and Enrichment from Diverse Linked
Data Sources
© Efstratios Kontopoulos © Panagiotis Mitzias © Marina Riga © Ioannis Kompatsiaris
Information Technologies Institute,
GR-57001 Thessaloniki, Greece
skontopo@iti.gr pmitzias@iti.gr mriga@iti.gr ikom@iti.gr
Abstract. Ontologies are a rapidly emerging paradigm for knowledge representation, with a growing
number of applications in data-intensive domains. However, populating enterprise-level ontologies with
massive volumes of data is a non-trivial and laborious task. Towards tackling this problem, the field of
ontology population offers a multitude of approaches for populating ontologies with instances in an automated
or semi-automated way. Nevertheless, most of the related tools typically analyse natural language text and
neglect more structured types of information like Linked Data. The paper argues that the rapidly increasing
array of published Linked Datasets can serve as the input for large-scale ontology population in data-intensive
domains and presents PROPheT, a novel software tool for ontology population and enrichment. PROPheT
can populate a local ontology model with instances retrieved from diverse Linked Data sources served by
SPARQL endpoints. As demonstrated in the paper, the tool is domain-agnostic and can efficiently handle vast
volumes of input data. To the best of our knowledge, no existing tool can offer PROPheT’s diverse extent of
functionality.
queries. Linked Data are formalised using controlled
1 Introduction vocabulary terms based on ontologies and can be
Ontologies constitute a knowledge representation publicly accessible via a SPARQL endpoint [4].
paradigm for modelling domains, concepts and This paper argues that the rapidly increasing array of
interrelations, effectively enabling the sharing of published Linked Datasets [1] can serve as the input for
information between diverse systems [23].The rapidly large-scale ontology population in DIDs and presents
emerging popularity of ontologies has led to their PROPheT, a software tool for user-driven ontology
deployment in various Data Intensive Domains (DIDs), population from Linked Data sources. The tool is
like e. g. bioinformatics [7], e-commerce [11] and digital domain-agnostic and can efficiently handle vast volumes
libraries [3]. Nevertheless, in order for ontologies to be of input data. To the best of our knowledge, no existing
further used at an enterprise level, massive volumes of tool can offer PROPheT’s extent of functionality.
data are required for populating the underlying models. The rest of the paper is structured as follows: Section
If performed manually, this task is extremely time- 2 gives an overview of related work approaches. Section
consuming and error-prone. Ontology population 3 presents PROPheT in detail, followed by a discussion
attempts to alleviate this problem, by introducing on PROPheT’s performance with regards to key
methods and tools for automatically augmenting an challenges for accessing information served by SPARQL
ontology with instances of concepts and properties. The endpoints. Section 5 presents an illustrative use case that
schema of the ontology itself is not altered but only its demonstrates the tool’s versatility and scalability.
set of concepts and relations. This process is part of Section 6 presents an evaluation of PROPheT, and the
ontology learning, which refers to the automatic (or paper is concluded with final remarks and directions for
semi-automatic) construction, enrichment and adaptation future work.
of ontologies [16].
2 Related Work
The vast majority of ontology population tools and
methodologies are aimed at textual input, typically Ontology population has already been deployed in
extracting knowledge from natural language text [5], various domains, like e.g. e-tourism [22], web services
[20]. However, other more structured sources of [21] and clinical data [17], amongst others. Regarding the
information are very often neglected. Such an example is application of ontology population in DIDs, we only
Linked Data [10], which builds upon standard Web came across a recent (2016) work by Knoell et al.
technologies and is a standard for publishing interlinked revolving around Big Data [15], indicating a potentially
structured data that are capable of responding to semantic emerging interest in the area.
Overall, and as already mentioned in the introduction,
Proceedings of the XIX International Conference state-of-the-art ontology population approaches in
“Data Analytics and Management in Data Intensive literature are mostly addressed to retrieving instances
Domains” (DAMDID/RCDL’2017), Moscow, Russia, from textual corpora (i.e. natural language text, like e.g.
October 10–13, 2017
184
product catalogues) and mainly involve machine enrichment from Linked Data sources in virtually any
learning, text mining and natural language processing domain, data-intensive or not.
(NLP) techniques. Other indicative approaches besides
3.2 Ontology Population
the ones discussed above are presented in [5] and [20].
A less popular stream of ontology population PROPheT offers the capability of class-based and
research is aimed at retrieving instances from other types instance-based ontology population. The former method,
of content, like e.g. CAD files [8], or more structured class-based population, retrieves instances from an
content, like e.g. spreadsheets [9], [13], and XML files external model and inserts them into a local ontology,
[19]. However, to the best of our knowledge, no other based on a class name entered by the user. Since the exact
approach similar to PROPheT currently exists that is class name has to be entered (e.g. dbo:Artist for the
capable of populating an ontology with instances DBpedia class representing artists), this method has the
retrieved from Linked Data sources, rendering PROPheT peculiarity that the user needs to know the structure of
into a highly novel tool. the external ontology. PROPheT then submits
appropriate SPARQL queries to the remote model’s
3 The PROPheT Ontology Population Tool endpoint and retrieves a result set of instances belonging
PROPheT1 is a novel software tool for ontology to the specified class. The user may then select the
population and enrichment that can retrieve instantiations instances to populate an existing class in the local
of concepts from SPARQL-served Linked Data sources ontology.
in a scalable manner. The retrieved instances are filtered The second method, instance-based population, has
based on user preferences and are then inserted into a two different modes:
target ontology. As described in the following (a) Retrieval based on instance label, which is performed
subsections, PROPheT provides various modes of according to a label (rdfs:label property value)
instance retrieval, along with the capability for entered by the user. The match of the retrieved
establishing user-defined mappings of the respective instances is based on an exact or partial match of the
properties. The tool’s mode of operation is purely user- input text.
driven, but relies on a step-by-step wizard-based (b) Retrieval based on an existing instance, in which the
interaction with the end-user, which greatly facilitates user selects an instance already existing in the local
use of the software even by largely unfamiliarised users. ontology and PROPheT queries the endpoint for
similar instances. More specifically, the tool finds
PROPheT’s front-end (see main window in Figure 1) classes in the remote ontology that include an
relies on Python and the PyQt application framework, instance with a similar rdfs:label property value
while the back-end deploys RDFLib and with the input instance. The user may then select
SPARQLWrapper, two Python APIs for manipulating specific classes, view their extension (i.e. set of
ontologies, along with an SQLite data store for storing instances) and choose which instances to import into
settings and user preferences. the local ontology model.
PROPheT is fully domain-independent in the sense In all the cases described above, and after the set of
that it can operate with any OWL ontology and any RDF preferred instances has been selected by the user to be
Linked Dataset that is served via a SPARQL endpoint. populated into the ontology, PROPheTlaunches the
3.1 Motivation ontology mapping process described next.
PROPheT was developed within the recently finished 3.3 Ontology Mapping
PERICLES FP7 project on Digital Preservation2. One of In order for PROPheT to proceed with the ontology
the domains tackled by the project was cultural heritage, population with the selected instances, a user-driven
where we faced the non-trivial challenge of populating ontology mapping is performed, in the sense that the
our domain ontologies with thousands of artefacts, each properties of the retrieved instances have to be mapped
of which was associated with hundreds of metadata to properties defined in the local model. In this context,
entries. In this affair, PROPheT was successfully PROPheT displays to the user a list of all datatype
deployed for populating the ontologies with instances properties (owl:DatatypeProperty) for the selected
retrieved from various Linked Data sources, like instances, in order for him/her to manually define
DBpedia and Freebase. appropriate mappings to datatype properties already
Nevertheless, though highly relevant [26], cultural existing in the local ontology. This mapping between
heritage is not the only DID where populating ontologies local and remote properties is mandatory for the property
from diverse sources poses a formidable challenge. Other values to be inserted into the local ontology along with
domains share similar concerns, like e.g. the the instances. For example, the user might define that the
telecommunications and news industry [2], and health retrieved property dbo:birthDate corresponds to the
and biomedicine [6], [14]. This was our main motivation local property ex:dateOfBirth. Once defined by the
for turning PROPheT into a truly domain-agnostic tool, user, PROPheT stores the mappings and offers
capable of performing ontology population and suggestions when the same mappings occur again
1 2
PROPheT is available at: http://mklab.iti.gr/project/prophet- http://www.pericles-project.eu/
ontology-populator
185
Figure 1 PROPheT’s main window
3.4 Instance Enrichment
Besides the ontology population capabilities • Discoverability, referring to how an endpoint can be
described above, PROPheT also offers the option of located and what are the available metadata;
enriching instances already existing in the local ontology • Interoperability, with regards to the supported
with properties and values from “similar” instances in SPARQL version(s);
remote ontologies; instance similarity here refers to
similarity in the respective instance labels (i.e. • Efficiency, which relates to the time needed to
rdfs:label). respond to the query;
The similar instances may belong to one or more • Reliability, based on the uptime of the endpoint on a
different classes in the remote ontology, thus, the tool constant basis.
presents the user with the type (rdf:type property A useful tool for monitoring the above parameters of
declaration) of each instance. Based on the content and SPARQL endpoints is SPARQLES [24], while the recent
semantics of the derived instances, the user may then Linked Data Fragments (LDF) paradigm promises to
decide which property-value pairs he/she will insert from alleviate the burden from endpoints, by redistributing the
the remote into the local ontology. load between clients and servers [25].
3.5 Ontology Enrichment Taking the above challenges into consideration, and
in order to demonstrate PROPheT’s scalability, we
The local model may also be semantically enriched experimented with timing the retrieval and population of
by establishing links between properties in the local and instances from the following well-known SPARQL
the remote ontologies via owl:equivalentProperty endpoints into a local custom ontology model:
declarations added into the local model. Similar links • DBpedia3, the Linked Data version of WikiPedia;
between classes are represented via owl:sameAs and/or
• OpenDataCommunities4, the official Linked Data
rdfs:seeAlso declarations added to the local ontology.
platform of the UK Department for Communities and
4 Challenges and PROPheT’s Performance Local Government (DCLG) that provides a selection
of official statistics and data outputs on a variety of
The availability and scalability of the SPARQL themes related to DCLG;
endpoints serving Linked Data is not always guaranteed,
• DBLP5, which provides open bibliographic
since maintaining such heavyweight query services
information on major computer science journals and
implies significant server-side costs coupled with various
proceedings;
potential technical problems on the level of the
infrastructure itself [4]. Key parameters for evaluating a • The Nobel Prize Linked Data dataset6 that contains
SPARQL endpoint are [4]: the authoritative information about Nobel prizes and
3 5
http://dbpedia.org/sparql http://dblp.l3s.de/d2r/sparql
4 6
http://opendatacommunities.org/sparql http://data.linkedmdb.org/sparql
186
hasPostalCode, etc.). This schema will be loaded in
Nobel Laureates since 1901; PROPheT to be populated.
• Eurostat statistics7 converted to RDF and re- Next, the user will need to register the sources that
published using Linked Data principles. serve the desired data (SPARQL endpoint URIs). For the
Table 1 illustrates the resulting retrieval and domain of the specific use case, there are several
population times for all selected endpoints. PROPheT’s established SPARQL-served ontologies that
performance is impacted by three parameters: (a) the contain instances of cities and towns, such as
software’s efficiency in querying and handling data, (b) ENVO8, an ontology of environmental features and
the endpoints’ speed in serving the requested data, and habitats, and LinkedGeoData9. Specifically,
(c) the volume of data (in the form of datatype property ENVO’s class City (ENVO_00000856) and
values) that the retrieved instances are attached to. LinkedGeoData’s classes City and Town contain
related instances.
Table 1 Instance retrieval and population times
Table 2 Instance retrieval and population times
No of Retrieval Population
Ontology Ontology No of instances Population time (sec)
instances time (sec) time (sec)
10 6,0 2,5 LinkedGeoData 10,000 120
100 19,0 8,3 ENVO 10,000 204
DBpedia LinkedGeoData 10,000 158
1,000 171,0 54,0
10,000 648,0 250,0
10 4,5 3,0 Taking advantage of PROPheT’s class-based
Open Data 100 18,0 6,7 instance extraction wizard, the user can respectively
Communities 1,000 104,0 44,0 populate two (or more) different classes of the local
10,000 510,0 210,0 schema with resources from two (or more) data sources.
10 3,5 1,8 For the purposes of this case study, PROPheT flawlessly
100 10,0 5,0 managed to retrieve and populate more than 30K
DBLP
1,000 62,0 32,0 instances, along with data property values. Specifically,
10,000 316,0 192,0 10K instances from ENVO’s City and 10K instances
10 3,7 2,0 from LinkedGeoData’s City were populated in the local
Nobel Prize
100 10,0 5,7 model’s class City. Also, another 10K instances from
1,000 56,0 31,0 LinkedGeoData’s class Town were imported to the local
10,000 270,0 170,0 model’s Town. Indicatively, Table 2 displays the
10 5,0 2,5 population times (in seconds) for the instances
100 15,7 7,7 mentioned above. Population times in the second batch
Eurostat
1,000 92,0 48,0 of LinkedGeoData instances is slightly higher, since the
10,000 440,0 225,0
local ontology already contained 20K instances
populated during the previous two phases.
Since parameter (a) remains constant within the
experiments, it becomes obvious that any differentiation Alternatively, supposing that the user cannot
in times heavily depends on parameters (b) and (c). predefine the classes of interest in the external models, a
Considering the facts that DBpedia reportedly contains different course will be followed. First, a single instance
the largest volume of property values, that most of the of the desired set will be located and imported. For
rest endpoints had almost equal number of properties and example, using PROPheT’s feature “Search by instance
that Eurostat’s selected instances had no datatype label”, the user will find a certain city of interest, e.g.
properties, it is clear that an endpoint’s response time Amsterdam, and import it into the local model. Next,
(second parameter) has a great impact on the ontology with the use of “Search for similar instances”, the
population process from Linked Data sources. software will discover all the classes where Amsterdam
is assigned to. Browsing the resulting list of classes, the
5 A Use Case Scenario user will now locate the classes of interest (e.g. class
City) and select more instances to be populated.
This section intends to demonstrate PROPheT’s Consequently, by utilizing the “Enrich Instance”
functionality by presenting a use case scenario in a data- function, the user can semantically enrich the major
intensive domain. Thus, consider a government cities’ instances (e.g. London, Paris, Amsterdam) with
institution monitoring pollution in rural environments, data regarding air pollution levels, residing in different
which requires a directory of cities and towns worldwide, endpoints.
enriched with related information, such as population, To conclude, the aforementioned use case
postal codes, etc. demonstrates PROPheT’s ability to populate various
Initially, a local ontology schema needs to be classes of an ontology with data retrieved from more than
deployed, incorporating the necessary classes (e.g. Town, one endpoints
City, etc.) and properties (e.g. hasPopulation,
7 9
http://eurostat.linked-statistics.org/sparql http://linkedgeodata.org
8
http://www.obofoundry.org/ontology/envo.html
187
.
Figure 2 Use case diagrammatic overview
Figure 2 illustrates a diagrammatic overview of the system’s initial requirements in terms of resources or
use case described in this section. The population-related background knowledge. PROPheT’s only requirement is
features permit different approaches for searching and that a local OWL ontology is already available, in order
browsing the available information, offering, thus, great to be populated with objects retrieved from Linked Data
flexibility to the user. sources. No domain-dependant resources are needed,
since PROPheT can flexibly adapt to any thematic
6 PROPheT Evaluation domain. No specialised software should be installed in
We recently conducted a user evaluation on the host machine either; PROPheT is distributed as a
PROPheT with very positive results [18]. As indicated standalone bundle.
by the resulting evaluation of the participants, the Learning approach: Refers to the system’s approach
following aspects of the tool were the most positive ones: in extracting knowledge and whether this approach is
attractiveness (93.5%), user-friendliness (93.5%), ease of specialised to a domain. Ontology population tools
usage (100%), innovativeness (87.5%), and efficiency typically employ Machine Learning techniques (see
(93.5%); the numbers in parentheses correspond to the Section 2), via statistical methods to identify terms or via
respective percentages indicating acceptance on behalf of automated pattern extraction. PROPheT, on the other
the users. The current section now presents a qualitative hand, deploys a purely user-driven, step-by-step
evaluation of the tool, based on the categorisation criteria ontology population and enrichment approach, which is
for ontology population tools proposed in [20]. suitable even for users with only fundamental familiarity
Elements extracted: Refers to the capacity of an with the pertinent notions.
ontology populating system to extract the various Degree of automation: A fully automated ontology
ontological aspects, like e.g. objects and relations. population system is of course desirable, but it seldom is
PROPheT offers the capability of extracting from possible to achieve, as the involvement of a domain
external sources both objects (i. e. class instances) and expert or an ontology engineer is very often needed. The
relations (i.e. data property values), and inserts them into PROPheT approach is mainly user-driven, requiring the
a local ontology model. Additionally, PROPheT also involvement of an end user for performing ontology
appends properties for semantically enriching the local population and enrichment through a step-by-step
model via owl:equivalentProperty, owl:sameAs wizard-based graphical user interface. Thus, although the
and rdfs:seeAlso. tool requires user intervention at each step, the process is
Initial requirements: This criterion refers to the achieved in a highly user-friendly fashion, as
188
demonstrated by our recent user evaluation of the tool Nevertheless, there are still a few areas of
[18] that indicated very positive feedback on behalf of improvement for the tool. In its current implementation,
the users. PROPheT is only limited to handling datatype and not
Consistency maintenance and redundancy object properties; the latter are significantly more
elimination: This criterion refers to the system's complex to tackle. Additionally, the tool cannot currently
capability to maintain the consistency of the ontology, handle direct or indirect imports of ontologies. A further
which is highly crucial, and to reduce redundancy, which improvement could be considering additional semantic
is not equally vital but can facilitate the process of enrichment associations, like e.g. skos:narrower and
querying the ontology and can limit its size and skos:broader from SKOS [12]. And, finally, the
complexity. Consistency maintenance in PROPheT is process of suggesting similar instances or classes to the
ensured by the integrated specialised APIs for user during the population and enrichment steps could be
manipulating ontologies and SPARQL queries. On the suggested by the tool itself, according to appropriate
other hand, the problem of instance redundancy (i.e. two similarity metrics. We are currently working on a revised
or more instances in the ontology referring to the same version of the software, which will integrate the
real object) is handled by PROPheT in a way that improvements mentioned above.
instances with the same name-identifier cannot be
populated multiple times in the ontology, i.e. values of
Acknowledgements
populated data properties are linked to one single This research received funding by the European
instance. Moreover, we are currently investigating Commission Seventh Framework Programme under
adding more complex handling mechanisms, such as Grant Agreement Number FP7-601138 PERICLES. We
heuristics or machine learning methods to identify would also like to thank the anonymous reviewers for
similar resources. their valuable remarks, thanks to which the paper has
Domain portability: This is an important aspect for been significantly improved.
all ontology populating systems and refers to their
capability to be ported to multiple thematic domains or
References
not. PROPheT is a totally domain-agnostic tool that is [1] Abele, A., McCrae, J.P., Buitelaar, P., Jentzsch, A.,
able to operate equally successfully in any domain, as Cyganiak, R.: Linking Open Data Cloud Diagram
long as the external sources are served through SPARQL (2017). http://lod-cloud.net/.
endpoints. Thus, no domain-specific knowledge is [2] Belam, M.: What is the Value of Linked Data to the
incorporated into the system. News Industry? The Guardian (2010, January).
Corpora modality: A system that is able to process https://www.theguardian.com/help/insideguardian/2
various modalities demonstrates its ability to 010/jan/25/news-linked-data-summit
accommodate and exploit diverse knowledge sources. In [3] Buckingham Shum, S., Motta, E., Domingue, J.:
this context, PROPheT can only process input from ScholOnto: An Ontology-based Digital Library
Linked Data sources through SPARQL endpoints, but Server for Research Documents and Discourse.
can easily be extended to process third-party ontologies International J. on Digital Libraries, 3 (3), pp. 237-
as well, retrieving instances and enriching the local 248 (2000)
ontology with additional property values found in these [4] Buil-Aranda, C., Hogan, A., Umbrich, J., Van-
models. denbussche, P.Y.: Sparql Web-querying Infra-
7 Conclusions and Future Work structure: Ready for Action? Int. Semantic Web
Conf., pp. 277-293. Springer (2013)
The paper argued that, with the rapidly emerging [5] Buitelaar, P., Cimiano, P.: Ontology Learning and
advent of the use of ontologies in data-intensive domains, Population: Bridging the Gap Between Text and
the process of ontology population becomes increasingly Knowledge, 167, Ios Press (2008)
relevant. Most proposed solutions are typically aimed at
[6] Callahan, A., Cruz-Toledo, J., Dumontier, M.:
analysing natural language text, often overlooking other
Ontology-based Querying with Bio2RDF’s Linked
sources of more structured information, like e.g. Linked Open Data. J. of Biomedical Semantics, 4 (1), S1
Data. In this context, we presented PROPheT, a domain (2013)
independent software tool for ontology population and
enrichment from Linked Data sources. Through wizard- [7] Gene Ontology Consortium. The Gene Ontology
based user-driven processes, the tool facilitates the semi- (GO) Database and Informatics Resource. Nucleic
automatic retrieval of instances and their insertion into a Acids Research, 32 (suppl 1), D258-D261 (2004)
local OWLontology. An advanced mapping process [8] Häfner, P., Häfner, V., Wicaksono, H.,
enables the dynamic definition of matching classes and Ovtcharova, J. Semi-automated Ontology Population
properties between source and target models. from Building Construction Drawings. KEOD, pp.
PROPheT’s rich functionality and versatility cannot be 379-386 (2013)
matched by any other ontology population tool found in [9] Han, L., Finin, T., Parr, C., Sachs, J., Joshi, A.:
literature, making PROPheT a truly innovative system RDF123: From Spreadsheets to RDF. Proc. of the 7th
for populating and enriching ontologies. Int. Semantic Web Conf. pp. 451-466, Springer
(2008)
189
[10] Heath, T., Bizer, C.: Linked Data: Evolving the Web the Semantic Web, pp. 31-41. Springer International
into a Global Data Space. Synthesis Lectures on the Publishing (2016)
Semantic Web: Theory and Technology, 1 (1), pp. 1- [19] Modica, G., Gal, A., Jamil, H.M.: The Use of
136 (2011) Machine-Generated Ontologies in Dynamic
[11] Hepp, M.: Goodrelations: An Ontology for Information Seeking. Cooperative Information
Describing Products and Services Offers on the Web. Systems, pp. 433-447, Springer (2001)
Int. Conf. on Knowledge Engineering and [20] Petasis, G., Karkaletsis, V., Paliouras, G.,
Knowledge Management, pp. 329-346. Springer Krithara, A., Zavitsanos, E.: Ontology Population
(2008) and Enrichment: State of the Art. Knowledge-driven
[12] Isaac, A., Summers, E.: SKOS Simple Knowledge Multimedia Information Extraction and Ontology
Organization System. Primer, World Wide Web Evolution, pp. 134-166. Springer-Verlag (2011)
Consortium (W3C) (2009) [21] Reyes-Ortiz, J.A., Bravo, M., Pablo, H.: Web
[13] Jupp, S., Horridge, M., Iannone, L., Klein, J., Owen, Services Ontology Population through Text
S., Schanstra, J., … Stevens, R.: Populous: A Tool for Classification. Computer Science and Information
Building OWL Ontologies from Templates. BMC Systems (FedCSIS), 2016 Federated Conference,
Bioinformatics, 13 (Suppl 1), S5 (2012). IEEE, pp. 491-495 (2016)
http://doi.org/10.1186/1471-2105-13-S1-S5. [22] Ruiz-Martınez, J.M., Minarro-Giménez, J.A.,
[14] Jupp, S., Malone, J., Bolleman, J., Brandizi, M., Castellanos-Nieves, D., Garcıa-Sánchez, F.,
Davies, M., Garcia, L., ... Wimalaratne, S.M.: The Valencia-Garcia, R.: Ontology Population: An
EBI RDF Platform: Linked Open Data for the Life Application for the E-tourism Domain. Int. J. of
Sciences. Bioinformatics, 30 (9), pp. 1338-1339 Innovative Computing, Information and Control
(2014) (IJICIC), 7 (11), pp. 6115-6134 (2011)
[15] Knoell, D., Atzmueller, M., Rieder, C., Scherer, K.P.: [23] Uschold, M., Gruninger, M.: Ontologies: Principles,
BISHOP-Big Data Driven Self-Learning Support for Methods and Applications. The Knowledge
High-performance Ontology Population. LWDA, pp. Engineering Review, 11 (02), pp. 93-136 (1996)
157-164 (2016) [24] Vandenbussche, P.Y., Umbrich, J., Matteis, L.,
[16] Maedche, A., Staab, S.: Ontology Learning from the Hogan, A., Buil-Aranda, C.: SPARQLES:
Semantic Web. IEEE Intelligent Systems, 16 (2), pp. Monitoring Public SPARQL Endpoints. Semantic
72-79 (2001) Web (Preprint), pp. 1-17 (2016)
[17] Mendes, D., Rodrigues, I.P., Baeta, C.F.: [25] Verborgh, R., Vander Sande, M., Hartig, O., Van
Development and Population of an Elaborate Formal Herwegen, J., De Vocht, L., De Meester, B., ...,
Ontology for Clinical Practice Knowledge Colpaert, P.: Triple Pattern Fragments: A Low-cost
Representation. KEOD, pp. 286-292 (2013) Knowledge Graph Interface for the Web. Web
[18] Mitzias, P., Riga, M., Kontopoulos, E., Stavropoulos, Semantics: Science, Services and Agents on the
T.G., Andreadis, S., Meditskos, G., Kompatsiaris, I. World Wide Web, 37, pp.184-206 (2016)
User-Driven Ontology Population from Linked Data [26] Wacker, M.: Linked Data for Cultural Heritage,
Sources. Int. Conf. on Knowledge Engineering and edited by Ed Jones and Michele Seikel. Chicago:
ALA Editions (2016)
190