Exploiting Linked Open Data for Enhancing MediaWiki-based Semantic

Organizational Knowledge Bases

Matthias Frank

and Stefan Zander

FZI Research Center for Information Technology, Haid-und-Neu-Str. 10-14, Karlsruhe, Germany

Fachbereich Informatik, Hochschule Darmstadt, Darmstadt, Germany

Keywords:

Linked Open Data, Semantic Web, Wiki Systems, Knowledge Engineering.

Abstract:

One of the main driving forces for the integration of Semantic Media Wiki systems in corporate contexts

is their query construction capabilities on top of organization-speciﬁc vocabularies together with the possi-

bility to directly embed query results in wiki pages. However, exploiting knowledge from external sources

like other organizational knowledge bases or Linked Open Data as well as sharing knowledge in a meaning-

ful way is difﬁcult due to the lack of a common and shared schema deﬁnition. In this paper, we introduce

Linked Data Wiki (LD-Wiki), an approach that combines the power of Linked Open Vocabularies and Data

with established organizational semantic wiki systems for knowledge management. It supports suggestions

for annotations from Linked Open Data sources for organizational knowledge bases in order to enrich them

with background information from Linked Open Data. The inclusion of potentially uncertain, incomplete,

inconsistent or redundant Linked Open Data within an organization’s knowledge base poses the challenge of

interpreting such data correctly within the respective context. In our approach, we evaluate data provenance

information in order to handle data from heterogeneous internal and external sources adequately and provide

data consumers with the latest and best evaluated information according to a ranking system.

1 INTRODUCTION

The adoption of semantic wiki approaches in orga-

nizational contexts and corporate environments has

recently begun and is continuously growing (Ghi-

dini et al., 2008; Kleiner and Abecker, 2010; Aveiro

and Pinto, 2013). This is particularly the case

for Semantic MediaWiki (SMW) (Kr

otzsch et al.,

2006), an extension for the popular MediaWiki en-

gine of Wikipedia, which introduces elements of

the W3C’s semantic technology stack

(W3C, 2007)

such as the Resource Description Framework’s triple

model (Klyne and Carroll, 2004), semantic proper-

ties (so-called roles in Description Logic terms) as

well as Concepts, i.e., dynamic categories that resem-

ble the notion of domains in the RDF Schema lan-

guage (Brickley and Guha, 2004). Those semantic

features in conjunction with its collaborative know-

ledge engineering capabilities make semantic Media-

Wiki systems even more attractive for a deployment

in professional environments (cf. listing “Wiki of the

A newer version of the semantic Web technology can

be accessed at: https://smiy.wordpress.com/2011/01/10/

the-common-layered-semantic-web-technology-stack/

Month”

). SMW provides enhanced query construc-

tion capabilities with respect to organization-speciﬁc

vocabularies and their speciﬁc contexts and allows to

treat query results as ﬁrst-class citizens and present

them dynamically within wiki pages. Organizations

like enterprises, NGOs or civil services can beneﬁt

from such features, which enable query construction,

query expansion, and ﬁltering using a lightweight set

of ontological semantics (Vrandecic and Kr

otzsch,

2006; Zander et al., 2014).

However, although existing semantic wiki ap-

proaches like SMW, Ontowiki or Wikibase are built

upon established semantic Web technologies, their

utilization in wiki-based representation frameworks

is primarily bound to a syntactic level. Moreover,

those systems focus on building organization-speciﬁc

lightweight ontologies and do not incorporate a com-

mon schema knowledge (cf. (Janowicz et al., 2014))

and its semantics per default. As a consequence, cur-

rent semantic wiki systems are not able to exploit

and beneﬁt from the growing availability of Linked

https://www.semantic-mediawiki.org/wiki/

Wiki of the Month

Frank M. and Zander S.

Exploiting Linked Open Data for Enhancing MediaWiki-based Semantic Organizational Knowledge Bases.

DOI: 10.5220/0006587900980106

In Proceedings of the 9th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (KEOD 2017), pages 98-106

ISBN: 978-989-758-272-1

Open Data (LOD) (see (Bizer et al., 2009; Hausen-

blas, 2009; Heath and Bizer, 2011)). Moreover,

the exploitation of additional knowledge from exter-

nal sources hosted by other organizations or Linked

Open Data sources as well as sharing knowledge in

a meaningful way across organizational boundaries

is difﬁcult due to the lack of a common vocabu-

lary among these approaches. Figure 1 illustrates the

different levels of open data according to Berners-

Lee (Berners-Lee, 2009):

Figure 1: Levels of Open Data.

The proposed approach overcomes the limita-

tion of lacking schema knowledge in organizational

knowledge bases

by supporting the annotation of

organization-speciﬁc schema knowledge with the

common terminology of Linked Open Vocabular-

ies (Janowicz et al., 2014) and extend the schema

knowledge by interlinking modelled entities with en-

tities represented as Linked Open Data in the LOD

Cloud

. Based on the resulting extended and inter-

linked schema knowledge, the so-called TBox in de-

scription logics (Baader et al., 2003), we provide the

users with addition information for local entities

, so

that their correctness and validity can be evaluated

on the basis of acquired externally hosted data where

a common and shared agreement is prevalent. This

leads to the following research questions:

• RQ1: How can users of organizational wikis be

supported in establishing new links to Linked

Open Data entities?

• RQ2: How can provenance information related to

entities in an organizational wiki be represented,

especially if these statements are inferred or gath-

ered from Linked Open Data?

We use the terms ‘knowledge base’ and ‘semantic wiki’

interchangeably throughout this work.

The Linking Open Data cloud diagram: http://

lod-cloud.net/

I.e., data or facts hosted internally in a local knowledge

base.

• RQ3: How can potentially uncertain, incomplete,

inconsistent or redundant Linked Open Data be

identiﬁed and tracked in order to increase the in-

formative value of an organizational knowledge

base?

We hypothesise that the information value of or-

ganizational knowledge bases will increase with the

integration of LOD. For the evaluation of this hy-

pothesis, we test our approach with the existing or-

ganizational wiki of our research group and compare

the information derived from LOD with veriﬁed in-

formation wherever possible and calculate the rate of

correctly derived information in relation to the false

derivations. If this relation is better than the relation

of our existing wiki, we regard our hypothesis as con-

ﬁrmed for this speciﬁc use case.

The remainder of this work is organized as fol-

lows: In Section 2, we discuss current semantic wiki

approaches wrt. the implementation of semantic Web

technology both on a syntactic and semantic level. In

Section 3, we detail our approach of interlinking or-

ganizational knowledge bases with LOD. The imple-

mentation of the approach is described in Section 4.

We conclude the current state of our work in Section 6

and discuss future work.

2 RELATED WORK

The review of related works is separated in two parts:

In Section 2.1 existing semantic wiki software is re-

viewed where special emphasis is given to their open-

ness towards a semantic technology stack. The ﬁnd-

ings are then summarized in Section 2.2 and close this

section.

2.1 Semantic Wiki Software

Some software applications for creating semantic

wikis do already exist. One of the best know ap-

plication is SMW, see (Kr

otzsch et al., 2006). As

many other wiki approaches, SMW is based on the

MediaWiki engine, which is famous for providing the

technical base for Wikipedia. The latest release

Semantic MediaWiki does support the development

of an organization speciﬁc knowledge base and the

querying of this data within the wiki. It is also pos-

sible to export the semantically described facts to

an external RDF store which does also allow to use

SPARQL for querying the data. More extensions for

https://github.com/SemanticMediaWiki/

SemanticMediaWiki/releases

Table 1: Characteristics of Semantic Wiki Applications.

Approach Underlying Engine Internal Data

Storage

Data Export

Format

Query con-

struction

Integration of

LOV/LOD

Semantic MW MediaWiki Relational

(RDF mirror

possible)

RDF (OWL

only)

#ask:

(SPARQL)

manual import

of single terms

OntoWiki - Relational or

RDF

RDF SPARQL publish ontol-

ogy with LOV

WikiBase MediaWiki Relational JSON or RDF WB-Client

(SPARQL)

WikiData-

scheme

Cargo MediaWiki Relational CSV #cargo query

(SQL-like)

MediaWiki exist that provide better syntactical link-

age of data modelled in SMW and RDF data like the

Triple Store Connector (ontoprise GmbH, discontin-

ued), the SparqlExtension

, or the RDFIO

extension.

All these approaches have in common that they pro-

vide semantic web technology on a syntactical layer,

rather than a semantic web integration on a semantical

layer. Only the LinkedWiki

extension focuses on

exploiting LOD for organizational knowledge bases

which are built on top of the MediaWiki engine. Due

to the fact that most semantic wiki applications are

just used to structure and query data within an organi-

zational wiki, rather than integrating data on a seman-

tical level, Koren (Koren, 2015) presents the Media-

Wiki extension Cargo. The Cargo extension does also

provide functionality for structuring on querying data,

but without employing any semantic web technology.

Rather than using semantic web technology, Cargo

provides a wrapper around relational databases and

exploits the well-established functionality of SQL.

One example for a non-MediaWiki based semantic

wiki applications is OntoWiki, see (Auer et al., 2007).

OntoWiki has its focus on modelling a plain know-

ledge base, without providing a knowledge presenta-

tion for human readers like free text and natural lan-

guage. A summarization of the previously discussed

characteristics of semantic wiki applications is pre-

sented in Table 1.

Although the semantic wiki software applications

introduced in Section 2.1 support semantic web tech-

nology like the Resource Description Framework

(RDF) or even the SPARQL Protocol and RDF Query

Language (SPARQL) on a syntactical level, the data

integration across multiple data sources is still hard

due to a common data scheme on a semantical level.

Vrandecic and Kr

otzsch (Vrandecic and Kr

otzsch,

2014) describe the collaborative data scheme in Wiki-

https://www.mediawiki.org/wiki/

Extension:SparqlExtension

https://www.mediawiki.org/wiki/Extension:RDFIO

https://www.mediawiki.org/wiki/Extension:LinkedWiki

Data as one possible solution for a common data

scheme in order to extend schema knowledge in other

wikis, especially Wikipedia. However, this approach

does also deﬁne a data schema which is independent

from Linked Open Vocabulary (LOV). In contrast to

the WikiData approach, the OpenAnno approach by

Frank and Zander (Frank and Zander, 2016) is fo-

cused on mapping proprietary ontologies to LOV in

order to support the interlinkage of local knowledge

bases with existing LOD.

2.2 Discussion

In Section 2.1 we have shown that current semantic

wiki applications provide technical integration of se-

mantic Web technology on a syntactic level. How-

ever, the introduced semantic wiki approaches do not

support the annotation and interlinkage of organiza-

tional knowledge with LOD on a semantic level while

considering the formal, model-theoretic semantics of

the underlying ontology language, i.e., a vocabulary’s

formal semantics. Such a recommendation system is

provided by Open-Anno, but it is not integrated in

any of the introduced semantic wiki applications. The

statements maintained by one of these semantic wiki

applications cannot be updated by external services

as the statements contained within a wiki are always

considered as master data. When importing state-

ments from external sources into an organizational

wiki, none of the introduced semantic wiki applica-

tions consider the context or the linkage of the data.

Both is important in order to evaluate given state-

ments, especially when they are inconsistent, redun-

dant or ambiguous. To overcome these limitations,

we introduce our Linked Wiki approach in Section 3.

3 APPROACH

In this Section, we introduce the Linked Data Wiki

(LD-Wiki), our approach to combine the power of

Linked Open Vocabularies and Data with established

organizational semantic wiki approaches.

3.1 Architecture

Our approach is technically based on MediaWiki in

combination with an RDF-store. The main contribu-

tion is thus not on a technical layer, but aims at sup-

porting the schema integration on a semantical layer.

We provide a set of established LOV to encourage the

reuse of these vocabularies in organizational wikis.

The resulting organizational knowledge base using

LOV is the foundation for suggestions of annotations

from LOD. These annotations allow to enrich the or-

ganizational knowledge base with additional informa-

tion from LOD. In order to distinguish organization-

speciﬁc statements from statements gathered from

LOD, we track the provenance information of each

statement. The provenance information is stored us-

ing named graphs in the RDF-store, which extends the

default triple model consisting of subject, predicate

and object to quadruples, containing an ID for each

statement. This ID allows us to attach provenance in-

formation to each statement. Using the provenance

information, we can also handle uncertain or incon-

sistent data and provide the data consumers with the

latest and most suitable information. One character-

istic of our approach is the strict separation of state-

ments in the triple store maintained by our extension

and the non-semantic part of the wiki, like free text,

MediaWiki syntax and place-holders for the semantic

statements, which are still maintained by the Media-

Wiki engine. By separating the semantic statements

from the non-semantic part of the wiki, we avoid the

issue of syncing statements between the wiki and the

knowledge base. Additionally, we are able to main-

tain and curate the semantic statements outside of the

wiki without causing inconsistent data. This separa-

tion of semantic and non-semantic data is therefore

a prerequisite for the transparent integration of state-

ments from the wiki itself and external statements

from LOD.

Figure 2 shows the architecture of our approach

including the two layers for knowledge management

and human friendly presentation. Our Linked Data

Management Module (LDaMM) is the stand-alone

business logic module for the knowledge manage-

ment layer which queries LOD on demand, updates

the local RDF knowledge graph and serves the Media-

Wiki engine for a human friendly presentation of the

knowledge graph. Other useful features of LDaMM,

which are not implemented yet, are reasoning and rule

execution which help to curate the local knowledge

graph. For the business logic module of the human

friendly presentation layer we employ the MediaWiki

engine which relies on the knowledge provided by

LDaMM, rather than relying on an own knowledge

serialization as it is done by e.g. SMW. Avoiding re-

dundant management of knowledge, we ensure that

the organizational knowledge management is always

in a consistent state. However, the MediaWiki en-

gine has to provide addition information for a hu-

man friendly presentation like free text and markup

information which is stored in a separate relational

database provided by a MySQL instance for local data

management on the presentation layer.

3.2 Challenges

In contrast to providing only wiki-based statements

within an organizational wiki, our approach does

also include external statements from multiple LOD

sources. The inclusion of external statements causes

issues when the same entity is described in multiple

sources. One of these issues is the fact of potential re-

dundant or inconsistent data. We address this issue by

exploiting the gathered provenance information and

evaluate the statements bases on a ranking derived

from contained provenance statements. The ranking

is inﬂuenced by the inter-linkage of the source as an

indicator of reference and by the evaluation of state-

ments by users of the wiki. Another issue is the poten-

tial amount of provenance information. Although this

provenance information is necessary in order to evalu-

ate the trustworthiness of statements, it would be con-

fusing for the user to show all available provenance

information for each statement. We address this chal-

lenge by evaluation the provenance information in the

background and just showing the resulting statement

to the user with an option to expand the underlying

provenance-based derivation of the statement.

4 IMPLEMENTATION

For the implementation of the LD-Wiki approach, we

build on the open source framework of MediaWiki.

In contrast to other MediaWiki-based approaches, we

implement the knowledge management as a stand-

alone module that controls storing, querying, updat-

ing, reasoning and rule execution of RDF-statements,

rather than integrating the knowledge management

in MediaWiki itself. This allows for a lightweight

MediaWiki extension that triggers the knowledge

management module for rendering wiki pages (Sec-

tion 4.1) and also if a user creates new wiki pages

for Terminological Box (TBox) (Section 4.2) or As-

sertional Box (ABox) (Section 4.3) of our knowledge

Linked Data Wiki

KnowledgePresentation

LocalWeb Business Logic

Quadruples

Store, Load, Update

Linked Open Data

Consume & Publish

LDaMM

Linked Data

Management Module

Updating, Querying,

Reasoning, Linking,

Rule Execution

Quadruples

Store, Load, Update

Free text. placeholder,

MW-syntax

Figure 2: Architecture of the Linked Open Data Wiki Approach: Knowledge layer for enriching the local RDF-graph with

Web knowledge and curating it, presentation layer for a user friendly presentation of RDF-data enriched with free text and

formatted by MediaWiki-Syntax stored in local MySQL-DB.

base.

4.1 Rendering Wiki Pages

The Wiki pages in our approach consist of free text

for a human readable presentation, placeholder for

data from the knowledge management module and

MediaWiki syntax to format the style of the page.

When a page is requested, the according parser func-

tion

of the LD-Wiki Extension requests the neces-

sary data from the knowledge management module

and replaces each place holder with the according

value from the knowledge base.

4.2 New TBox Pages

The key factor to let the LD-Wiki approach work well

and build a TBox which can be interpreted in the con-

text of LOD, it is necessary to interlink new concepts

with concepts from LOV. Concepts are represented as

categories in MediaWiki. Therefore, whenever a new

category is created, LDaMM is triggered to query for

https://www.mediawiki.org/wiki/Parser functions

existing concepts in LOV with the same label as the

label for the new category. If one or more classes are

found, the user of the wiki can select the concepts that

represent the intended meaning.

Figure 3 shows how this looks like in the LD-

Wiki. For creating a new concept within the local

knowledge management, the user opens the special

page for creating a new category in MediaWiki and

provides the string that labels that new concept.

When submitting this string, MediaWiki sends it

to LDaMM in the knowledge management layer.

LDaMM invokes SPARQL queries to search for

concepts in LOD that are labeled with the same

string. If, for example, the user would like to create a

new concept for cities for a German-language termi-

nology, he would probably enter the string ”Stadt” for

this concept. To ﬁnd concepts related to that string in

LOD, LDaMM produces the query string ’SELECT

* WHERE ?category rdf:type rdf:Class; rdfs:label

”Stadt”@de. ’ to discover any concept that has the

label ”Stadt” with a German language tag. This query

string is then executed at available public SPARQL

endpoints to discover adequate concepts. Expected

Figure 3: Interlink new category with existing concepts.

results would be for example http://schema.org/City,

http://dbpedia.org/ontology/City or

http://www.wikidata.org/entity/Q515. LDaMM

returns these results to MediaWiki where the user

can select the adequate concepts. On creation of the

new category in MediaWiki including the interlinked

concepts, the information of the new category and the

linked concepts are send back to LDaMM and stored

to the local knowledge graph.

4.3 New ABox Pages

Assuming that the categories of the LD-Wiki are

linked to the according concepts in LOV as discussed

in Section 4.2, we can assist the user on creating new

instances for the ABox in the wiki. Instances of a con-

cept are represented as pages within the category that

represents that concept in the wiki. Therefore, when-

ever a page is created, the knowledge management

module is triggered to query for existing individuals

in LOD with the same label as the new page and the

same concept as the category of the new page. If one

or more individuals are found, the user of the wiki can

select the individuals that represent the same instance

as the the page. The great beneﬁt for this kind of in-

terlinkage is that we can query directly for properties

of these individuals in LOD or retrieve a summary

of entity data using entity summarization tools like

LinkSUM (Thalhammer et al., 2016).

Figure 4 shows how this is done in the LD-Wiki.

For creating a new instance within the local know-

ledge management, the user opens the special page

for creating new instances in MediaWiki, provides

the string that labels that new instance and selects

the category of which the new page should be an in-

stance of. When submitting this string, MediaWiki

again sends it to LDaMM in the knowledge man-

agement layer together with the identiﬁer of the se-

lected category. LDaMM invokes SPARQL queries

to search for instances in LOD that are labeled with

the same string and are instances of any of the con-

cepts that the given category is linked to. If, for ex-

ample, the user would like to create a new instance

of the category ”Stadt” for the German-language ter-

minology in our example, he would enter the name

of this city as string, e.g. ”Karlsruhe”, and select

the category ”Stadt” for it. To ﬁnd instances re-

lated to that string and category in LOD, LDaMM

produces the query string ’SELECT * WHERE ?in-

stance rdf:type http://www.wikidata.org/entity/Q515;

rdfs:label ”Karlsruhe”. ’ to discover any in-

stance that has the label ”Karlsruhe” and type

http://www.wikidata.org/entity/Q515, as this is one

of the concept which is linked to the category

”Stadt”. This query string is then executed at avail-

able public SPARQL endpoints to discover adequate

intances. An expected result would be for exam-

ple http://www.wikidata.org/entity/Q1040 which de-

scribes the German city in the state of Baden-

Wuerttemberg. LDaMM returns these results to

MediaWiki where the user can select the adequate in-

stance. On creation of the new instance in MediaWiki,

the information of the new instance and the linked in-

stances is send back to LDaMM and stored to the lo-

cal knowledge graph, including all properties that are

retrieved from the linked entity and also their prove-

nance information. All this information is now avail-

able in the local knowledge graph without further ac-

tion.

Figure 4: Interlink new instance with existing individual.

5 EXPERIMENTAL RESULTS

For a ﬁrst evaluation of our approach, we use the

SPARQL endpoints of DBpedia

and Wikidata

two instances of LOD resources. Due to the different

implementation of these endpoints, the query string

has to be mapped to meet the individual characteris-

tics.

5.1 New TBox Pages

The ﬁrst step is to run the query SELECT * WHERE

{?category rdf:type rdf:Class; rdfs:label

"Stadt"@de .} on the SPARQL endpoints of

DBpedia and Wikidata.

Wikidata

Wikidata uses the property http://www.wikidata.org/

prop/direct/P279 (subclass of) to describe subclasses

of other classes. We therefore map the property-value

pair “rdf:type rdf:Class” to this wikidata property

which results in the following query:

SELECT * WHERE {?category

<http://www.wikidata.org/prop/direct/P279>

?class ; rdfs:label "Stadt"@de .}

When executing this query at the SPARQL

endpoint of Wikidata, we receive two classes:

http://www.wikidata.org/entity/Q515 which de-

scribes a city as a large and permanent human settle-

http://dbpedia.org/sparql

https://query.wikidata.org

ment and http://www.wikidata.org/entity/Q15253706

which is the class for a more speciﬁc deﬁnition of a

city by country that holds the size of cities and towns

in Korea, Japan, the USA, China, North Korea and

France.

DBpedia

For DBpedia, we map the class of rdfs:Class to

owl:Class as DBpedia makes use of Web Ontology

Language (OWL) and the default conﬁguration of

this endpoint does not imply superclasses which

would include rdfs:Class as well. The result is the

following query:

SELECT * WHERE {?category rdf:type

owl:Class ; rdfs:label "Stadt"@de .}

When executing this query at the SPARQL

endpoint of DBpedia, we receive again two

classes: http://dbpedia.org/ontology/City and

http://dbpedia.org/ontology/Town.

5.2 New ABox Pages

Next, we test the retrieval of instance data for a given

concept. In our example, we want to execute the

query SELECT * WHERE {?instance rdf:type

<http://www.wikidata.org/entity/Q515>;

rdfs:label "Karlsruhe"@de. } on the SPARQL

endpoints of DBpedia and Wikidata.

Wikidata

Wikidata uses the property http://www.wikidata.org/

prop/direct/P31 (instance of) to indicate that an

instance belongs to a speciﬁc category. We therefore

map the property rdf:type to the Wikidata-speciﬁc

term:

SELECT * WHERE {?instance

<http://www.wikidata.org/prop/direct/P31>

<http://www.wikidata.org/entity/Q515> ;

rdfs:label "Karlsruhe"@de .}

For this query, we get two matching instances:

http://www.wikidata.org/entity/Q1040, the Ger-

man city in the state of Baden-Wuerttemberg, and

http://www.wikidata.org/entity/Q1026577, a city

in North Dacota. Depending on the instance the

user wants to refer to, he or she has to select the

appropriate one. This example does also show that a

completely automatic information retrieval is difﬁcult

to control and therefore human supervision of this

process is still reasonable. If we run the query

with the more deﬁnition of a city by country using

the query string SELECT * WHERE {?instance

<http://www.wikidata.org/prop/direct/P31>

<http://www.wikidata.org/entity/Q15253706>;

rdfs:label "Karlsruhe"@de . }, we do not get

any result.

DBpedia

For DBpedia, we run the query for instances

of http://dbpedia.org/ontology/ City or http://

dbpedia.org/ontology/Town:

SELECT * WHERE {?instance rdf:type

<http://dbpedia.org/ontology/Town> ;

rdfs:label "Karlsruhe"@de . } The sin-

gle result of this query is the instance of

http://dbpedia.org/resource/Karlsruhe.

6 CONCLUSION

With the LD-Wiki approach we have shown how we

can assist users of organizational wikis with creating

new links to LOD entities. As we have separated the

knowledge management module from the knowledge

representation in MediaWiki, we are able to keep

track of the provenance of statements in our know-

ledge base without affecting the knowledge represen-

tation. Especially the synchronization of wiki data

and the organizational knowledge base as it was the

case in other semantic wiki approaches like SMW is

not required any longer. The great beneﬁt for this kind

of interlinkage is that we can enrich the information

value of individuals in LD-Wiki by querying for prop-

erties of these individuals in LOD or retrieve a sum-

mary of entity data using entity summarization tools

to exploit the power the continuously growing amount

of LOD for corporate knowledge bases. Open issues

for future work include the privacy for conﬁdential

data on the one hand while publishing parts of the

corporate knowledge base as LOD on the other hand.

This requires a proper implementation of Access Con-

trol Lists (ACLs) with carefully designed access roles

for each statement in the knowledge base. In or-

der to beneﬁt from features of semantic Web tech-

nologies besides reusing information from LOD, the

knowledge management module should also enable

advanced reasoning over organizational data which

would also help to evaluate and interpret potential un-

certain, incomplete, inconsistent or redundant LOD

correctly. This would further increase the informative

value of the organizational knowledge base.

REFERENCES

Auer, S., Jungmann, B., and Sch

onefeld, F. (2007). Se-

mantic wiki representations for building an enterprise

knowledge base. In Reasoning Web, volume 4636 of

Lecture Notes in Computer Science, pages 330–333.

Springer Berlin Heidelberg, Berlin, Heidelberg.

Aveiro, D. and Pinto, D. (2013). Implementing organiza-

tional self awareness - a semantic mediawiki based

enterprise ontology management approach. In Filipe,

J. and Dietz, J. L. G., editors, KEOD, pages 453–461.

SciTePress.

Baader, F., Calvanese, D., McGuinness, D., Nardi, D., and

Patel-Schneider, P. (2003). The Description Logic

Handbook: Theory, Implementation and Applications.

Cambridge University Press.

Berners-Lee, T. (2009). Linked-data design

issues. W3C design issue document.

http://www.w3.org/DesignIssue/LinkedData.html.

Bizer, C., Heath, T., and Berners-Lee, T. (2009). Linked

data – the story so far. International Journal on Se-

mantic Web and Information Systems, 5(3):1–22.

Brickley, D. and Guha, R. (2004). RDF Vocabulary De-

scription Language 1.0: RDF Schema. W3C Recom-

mendation.

Frank, M. and Zander, S. (2016). Pushing the cidoc-

conceptual reference model towards lod by open an-

notations. In Oberweis, A. and Reussner, R. H., edi-

tors, Modellierung 2016, 2.-4. M

arz 2016, Karlsruhe,

LNI, pages 13–28. GI.

Ghidini, C., Rospocher, M., Seraﬁni, L., Kump, B., Pam-

mer, V., Faatz, A., Zinnen, A., Guss, J., and Lind-

staedt, S. (2008). Collaborative knowledge engineer-

ing via semantic mediawiki. In Auer, S., Schaffert,

S., and Pellegrini, T., editors, International Confer-

ence on Semantic Systems (I-SEMANTICS ’08), pages

134–141, Graz, Austria.

Hausenblas, M. (2009). Exploiting Linked Data to

Build Web Applications. IEEE Internet Computing,

13(4):68–73.

Heath, T. and Bizer, C. (2011). Linked Data: Evolving the

Web into a Global Data Space. Morgan & Claypool,

1. edition.

Janowicz, K., Hitzler, P., Adams, B., Kolas, D., and Varde-

man, C. (2014). Five stars of linked data vocabulary

use. Semantic Web, 5(3):173–176.

Kleiner, F. and Abecker, A. (2010). Semantic media-

wiki as an integration platform for it service manage-

ment. In F

ahnrich, K.-P. and Franczyk, B., editors, GI

Jahrestagung (2), volume 176 of LNI, pages 73–78.

GI.

Klyne, G. and Carroll, J. J. (2004). Resource description

framework (RDF): concepts and abstract syntax. W3C

recommendation, World Wide Web Consortium.

Koren, Y. (2015). Cargo and the future of semantic media-

wiki. In SMWCon Spring 2015. St. Louis, MO, USA.

otzsch, M., Vrande

c, D., and V

olkel, M. (2006). Se-

mantic mediawiki. In ISWC 2006, volume 4273 of

Lecture Notes in Computer Science, pages 935–942.

Springer Berlin Heidelberg, Berlin, Heidelberg.

Thalhammer, A., Lasierra, N., and Rettinger, A. (2016).

Linksum: Using link analysis to summarize en-

tity data. In Alessandro Bozzon, Philippe Cudr

Mauroux, C. P., editor, 16th International Conference

on Web Engineering (ICWE 2016), volume 9671 of

Lecture Notes in Computer Science, pages 244–261.

Springer.

Vrandecic, D. and Kr

otzsch, M. (2006). Reusing onto-

logical background knowledge in semantic wikis. In

olkel, M. and Schaffert, S., editors, SemWiki, vol-

ume 206 of CEUR Workshop Proceedings. CEUR-

WS.org.

Vrandecic, D. and Kr

otzsch, M. (2014). Wikidata: A

free collaborative knowledgebase. Commun. ACM,

57(10):78–85.

W3C (2007). Semantic web stack.

Zander, S., Swertz, C., Verd

u, E., P

erez, M. J. V., and

Henning, P. (2014). A semantic mediawiki-based ap-

proach for the collaborative development of pedagog-

ically meaningful learning content annotations. In

Molli, P., Breslin, J. G., and Vidal, M.-E., editors,

SWCS (LNCS Volume), volume 9507 of Lecture Notes

in Computer Science, pages 73–111. Springer.