=Paper=
{{Paper
|id=Vol-1890/paper05fullPaper
|storemode=property
|title=Porting the xEBR taxonomy to a linked open data compliant format
|pdfUrl=https://ceur-ws.org/Vol-1890/paper05fullPaper.pdf
|volume=Vol-1890
|authors=Thierry Declerck,Dagmar Gromann
}}
==Porting the xEBR taxonomy to a linked open data compliant format==
https://ceur-ws.org/Vol-1890/paper05fullPaper.pdf
Porting the xEBR Taxonomy to a Linked Open
Data compliant Format*
Thierry Declerck1 and Dagmar Gromann2
1
DFKI GmbH, Stuhlsatzenhausweg 3,
D-66123, Saarbrücken, Deutschland,
declerck@dfki.de,
https://www.dfki.de
2
Artificial Intelligence Research Institute (IIIA-CSIC), Campus de la UAB,
E-08193 Bellaterra, Spain,
dgromann@iiia.csic.es,
http://iiia-csic.es
Abstract. The XBRL Europe Business Registers (xEBR) working group
has developed a Core Reference Taxonomy over the last years. This work
represents a milestone in the support of conceptual interoperability con-
cerning the information on company identification and financial state-
ments across European legislations and languages. Our main contribu-
tion is to port the current data of this taxonomy, which is available both
as an Excel table and in the standard XML format of XBRL, to a Linked
Data compliant format in order to make the taxonomy interoperable also
at the semantic level. This paper describes the current version of the on-
tological model resulting from this transformation and explains some of
our design decisions.
Keywords: xEBR Taxonomy, Linked Open Data, Multilingualism
1 Introduction
Transparent financial reporting is key to preventing corporate fraud and allow-
ing for stakeholders to compare business results of different corporations over a
certain reporting period. However, business reports summarize several types of
information, such as financial statements and risk and performance reports, in
various formats for different jurisdictions and in specific natural languages. To
make this variety of information comparable, the eXtensible Business Reporting
Language (XBRL)1 has been introduced, a freely available and global business
reporting standard encoded in an XML-based language. This XML schema al-
lows for the tagging of specific items of company reports, which can then be
* We dedicate this paper to the memory of our colleague, Dr. Hans-Ulrich Krieger,
who recently passed away.
1
https://www.xbrl.org. See also [5].
automatically processed, such as being automatically compared in a spread-
sheet. Even though the focus of this XML-based format is on syntactic rather
than semantic correctness, its adoption by businesses has already shown to lead
to an increased business value as proven empirically (e.g. [8]). We believe that
this increase in value can be further tapped into by a semantically expressive
version of XBRL, especially for the European case, in which each member state
defines specific rules for filing business reports under its own jurisdiction, which
is reflected in its local XBRL taxonomy by means of different name-spaces and
tag names. Some tags only exist in specific local XBRL taxonomies.
To render those local XBRL taxonomies interoperable across borders, the
XBRL Europe Business Registers (xEBR) working group has developed the so-
called xEBR Core Reference Taxonomy2 . This taxonomy serves as an interlingua
between XBRL jurisdictions that represents shared information and assigns them
canonical names as well as further information, such as key financial ratios and
company identifiers. Several local XBRL taxonomies have already been aligned
with this centralized xEBR format, however, this alignment process is tedious
and its verification is difficult [13]. Furthermore, the XML-based xEBR format
solves primarily the syntactic interoperability problem, but does not provide a
semantically grounded representation across languages and jurisdictions [12].
Porting the xEBR Core Reference Taxonomy to a Linked Open Data (LOD)3
compliant format enables the use of semantic technologies. Thereby, continuously
updated financial information, such as stock exchange data, can be aligned with
more static data from XBRL reports (see e.g. [3]). Semantic technologies allow
for automated consistency checking of represented knowledge, which is bene-
ficial for the verification and generation of financial summaries. Thereby, the
comparison of companies reporting to different jurisdictions can be facilitated.
The LOD-compliant model we propose in this paper assigns in a declarative way
multilingual information to each concept, which eases its use by human analysts
across countries. By analysts we refer to all stakeholders of a company, including
shareholders, government, and also the general public.
This paper first sets the proposed work in relation to previous approaches
before we discuss the xEBR taxonomy in more detail. We then describe the
details of the transformation process from an Excel version of the taxonomy to
a machine-processable ontology in Section 4.
2 Related Work
The proposed approach addresses the problem of formalizing an existing financial
reporting taxonomy for the purpose of increasing the semantic interoperability
of its associated local taxonomies and to leverage the benefits of Semantic Web
technologies. To this end, we address two important problems that have been
2
See http://www.xbrleurope.org/working-groups/xebr-wg/xebr-taxonomy for
more details.
3
See http://linkeddata.org/ for more details.
tackled by other approaches before: a) alignment of xEBR and XBRL, and b)
formalization of accounting concepts.
2.1 xEBR and XBRL Alignments
To overcome the multiplicity in jurisdiction and natural language of reporting
standards, several approaches to align individual XBRL taxonomies have been
proposed. Spohr et al. [12] suggest a supervised ranking support vector machine
algorithm to perform multilingual ontology alignment. The system developed by
Spohr et al [12] is utilized as a baseline for a logic-based alignment approach
proposed by Thomas et al [13]. While there is a whole range of multilingual and
crosslingual ontology alignment approaches (see [14] for an overview), the two
described here are particularly relevant since they are also interested in expli-
cating the implicit semantics of XBRL taxonomies. However, the focus of this
paper is on the translation of an existing xEBR taxonomy to a LOD-compliant
format rather than formalizing alignments.
2.2 Formalizing Accounting Concepts
In the literature on formalizing accounting concepts the two major approaches
are directly specifying accounting principles in ontologies and transforming ex-
isting taxonomies to a Semantic Web format. In the first category, Bai et al. [1]
propose an ontology-based extension of XBRL to allow for its use in a financial
service matching framework. Thomas et al. [13] propose a logic-based approach
to explicate shared semantics that are implicit across XBRL taxonomies with
the objective to align those shared concepts across taxonomies from different ju-
risdictions. Krahel [7] suggests the use of a formalization process of accounting
principles to detect inconsistencies in existing standards.
Approaches to porting XBRL taxonomies to ontological formats [4, 2] and
to Linked Open Data [6, 10] are vital to linking XBRL data to other data on
the Web. In fact, a whole range of XBRL-related ontologies has been published,
an overview of which can be found in O’Rian [9]. While we base our research
on lessons learned from those previous approaches, most of them represent a
very literal conversion of the XBRL taxonomies with little semantic addition
[13]. More recent approaches suggested to port XBRL to multidimensional data
models in order to ensure the correctness of XBRL formulas [11]. While the dis-
cussion of format and standard in this work is very insightful, little is added to
the semantic explication of accounting principles and definitions in XBRL. In-
stead of a literal conversion, we focus on a semantic explication of the principles
represented in xEBR to unleash its full potential in an LOD-compliant format.
This also means closely analysing all existing definitions and representing them
in explicit semantics. In a second step this ontologization of xEBR facilitates
alignments across resources, however, this is not the focus of the current publi-
cation.
3 The xEBR Core Reference Taxonomy
The xEBR Core Reference Taxonomy defines common concepts for financial
statements and company identification, across a certain number of legislations
and countries4 . This way, matching tables have been proposed between the xEBR
taxonomy and local taxonomies, covering the legislations of Belgium, Germany,
France, Italy, The Netherlands, Spain and the United Kingdom. The xEBR
taxonomy also describes equivalences to codes of the BACH database5 .
Figure 1 below gives a very partial view on the xEBR taxonomy data encoded
in Excel6 .
Fig. 1. A screen shot of the Excel version of xEBR – preliminary version 8
In this partial view of the Excel sheet encoding the xEBR taxonomy, the
reader can see that the labels/terms used in the taxonomy have various (re-
current) endings that are carrying a specific meaning. In the second row of the
first column, for example, the term used is: “COMPANY BALANCE SHEET,
HORIZONTAL LAYOUT [REPORT]”, one of the eight different instantiations
of the type“role” (the various types the xEBR elements can belong to are listed
in the right column of the displayed partial Excel sheet). We observe that the
word used in square brackets has a special meaning: we are dealing with a “re-
port” element, which does not come as a surprise since XBRL is precisely about
4
Details on the xEBR taxonomy are given in http://www.xbrleurope.org/
working-groups/xebr-wg/xebr-taxonomy.
5
BACH stands for the ”Bank for the Accounts of Companies Harmonized”. See
https://www.bach.banque-france.fr/?lang=en for more details.
6
But the reader can download the full Excel table of a former version of the taxonomy
at http://www.xbrleurope.org/working-groups/xebr-wg/xebr-taxonomy
business reporting. The third row is introduced by the term “Assets [Presenta-
tion]”. This one is stating that “Assets” is a presentation element, which as such
is not associated with a specific value, but which is relevant for describing the
(conceptual) structure of a report.
As a first comment, we can state here that, contrary to the impression given
by the hierarchical organisation displayed in the first column of the Excel sheet,
the relation between a “report” element and a “presentation” element is not a
sub-class hierarchy but much more of a part-of relation. The presentation element
“Assets” is a (probably necessary) part of a XBRL report, to be included in its
specific “COMPANY BALANCE SHEET, HORIZONTAL LAYOUT” section.
While we can assume that there is a sub-class relation between the presentation
element “Assets” and the presentation element introduced in the fourth row of
the Excel sheet: “Subscribed capital unpaid [Presentation]”; our first interpre-
tation is that “Subscribed capital unpaid” is an “Asset”. We think that this
distinction between part-of and sub-class relations between reporting elements
listed in this hierarchical Excel structure is important.
We observe also that the information about the type of a term is included
in its typographical realisation. An exception for this are the terms that rep-
resent single financial monetary values (“Plant and machinery” vs. “Property,
plant, and equipment [Total]”, where “Total” clearly marks the associated type
“monetary”). This inclusion of the type name in the term is needed in order
to differentiate terms that would have the same typographical realisation, such
as “Property, plant, and equipment [Presentation]” vs. “Property, plant, and
equipment [Total]”. We believe that it is much better to avoid this strategy of
using this kind of typographical conventions for naming concepts, but instead to
use interpretation independent codes, which mark only the structural relations
“A.” and “A.1” followed by “A.1.2” or the like, as done in many taxonomies,
and to include the terms inside the objects associated with those codes in the
form of “labels”. Actually, standard XBRL practices include such labels in their
XML serializations, but keep the used labels for naming the concepts just by
leaving out the blank spaces and using a camel case notation (as can be seen
in the third column ”Element Name”) of the Excel sheet in Figure 1. We advo-
cate for replacing this convention by using more standard codes, in the form of
alpha-numeric sequences.
4 xEBR to Ontology
Our mapping from xEBR to a LOD-compliant format concerns first the hier-
archical organisation of the taxonomy, then the relations between the concepts
listed in the taxonomy and their types: “role”, “abstract”, “monetary”, “tuple”,
“string”, “boolean”, or “uri”, whereby “role” is subdivided (in this version of the
taxonomy) into eight classes (like ”CompanyBalanceSheetHorizontalLayoutRe-
port”, ”CompanyHistoryReport”, etc.). And we also foresee two sub classes for
the monetary role: “total” value and “single” item value. We finally address the
relations between the multilingual labels. In this paper we do not focus on the
information related to company identification, so that we will not expand on its
associated types “tuple”, “string”, “boolean” and “uri”.
Fig. 2. A screen shot showing partially the class hierarchy of the xEBR Ontology
In Figure 2 we display a partial view of the hierarchy of concepts we derived
from the xEBR taxonomy in an OWL and RDF environment. The reader can
see that we organize the xEBR concepts as sub-classes of the 4 main types:
“abstract”, “monetary”, “role” and “tuple”. The reader can also observe that
we are using alpha-numeric codes for encoding the original xEBR taxonomy
elements to be seen in the third column of the Excel sheet partially displayed in
Figure 2. The labels of the original xEBR elements are now represented by the
annotation property rdfs:label, as it is shown in Figure 3.
In the following we just present some examples of the way we ported elements
of xEBR to an OWL/RDF(S) ontology. The displayed examples are in the so-
called Turtle notation.
Fig. 3. A screen shot showing the encoding of the IntangibleAssetsPresentation element
The first example states that “AssetsPresentation” (encoded in our ontology
as P.1) is an owl:Class and as such a subClass of “Abstract”. The second example
repeats the same type of encoding, this time for the subClass “FixedAssetsPre-
sentation” (encoded as P.1.3 – showing the hierarchical relation to “AssetsPre-
sentation” by adding a“.” to“P.1”). The subClass property is transitive, so that
we do not have to state the partOf relation of P1.3 to R.1.
rdf:type owl:Class ;
xebr:isPartOf ;
xebr:version "V1" ;
rdfs:comment "Representation of the xEBR presentation element
AssetsPresentation"@en ;
rdfs:label "Assets [Presentation]"@en ;
rdfs:subClassOf xebr:Abstract ;
.
rdf:type owl:Class ;
xebr:version "V1" ;
rdfs:comment "Representation of the xEBR presentation element
FixedAssetsPresentation"@en ;
rdfs:label "Fixed assets [Presentation]"@en ;
rdfs:subClassOf ;
.
The partOf relation of “P.1” to the role “Report” in the first example above is
marked by the line“xebr:isPartOf ”. With R.1 we encode the original xEBR report element “CompanyBal-
anceSheetHorizontalLayoutReport”, as displayed just below:
rdf:type owl:Class ;
xebr:version "V1" ;
rdfs:comment "Representation of the xEBR report element
CompanyBalanceSheetHorizontalLayoutReport"@en ;
rdfs:label "COMPANY BALANCE SHEET, HORIZONTAL LAYOUT [REPORT]"@en ;
rdfs:subClassOf xebr:Role ;
.
The next example introduces another case of a “part-of” relation. The class
“FixedAssetsTotal”, which is of type “Total” (itself of type “Monetary”, and
therefore we encode it prefixing it with a “M”: M.1.3) is introduced as an element
that is part of the P.1.3 presentation element (it is not a subClass of this element,
as can be seen also in Figure 2, in which the Monetary has its own class hierarchy,
something new compared to the original xEBR taxonomy).
rdf:type owl:Class ;
xebr:isMonetaryPartOf ;
rdfs:comment "An element of type Monetary representing a total"@en ;
rdfs:label "Fixed Assets [Total]@en" ;
rdfs:subClassOf xebr:Total ;
.
The following example is also about an element that is of type “Monetary”,
but it is not representing a “Total” figure, therefore we introduced a “calculation”
property (very similar to what is foreseen in XBRL). This property is used to
relate the element “CostsOfDevelopment” (M.1.3.1.1) to the “Total” element it
contributes to: “IntangibleAssetsTotal” (M.1.3.1), which itself is a part of P.1.3.1
(“IntangibleAssetsPresentation”).
rdf:type owl:Class ;
xebr:isCalculationElementOf ;
rdfs:comment "Element of the class Monetary that
has a single value" ;
rdfs:label "Costs of development@en" ;
rdfs:subClassOf xebr:Single ;
.
rdf:type owl:Class ;
xebr:isMonetaryPartOf ;
rdfs:comment "An element of type Monetary representing a
total"@en ;
rdfs:label "Intangible assets [Total]@en" ;
rdfs:subClassOf xebr:Total ;
.
rdf:type owl:Class ;
xebr:version "V3" ;
rdfs:comment "Representation of the original xEBR element
IntangibleAssetsPresentation"@en ;
rdfs:label "Intangible assets [Presentation]"@en ;
rdfs:subClassOf ;
.
The following lists a few properties that we have created so far.
xebr:isCalculationElementOf
rdf:type owl:FunctionalProperty ;
rdf:type owl:ObjectProperty ;
rdfs:comment "This property marks an element as a part of a
total Monetary item" ;
rdfs:domain xebr:Monetary ;
rdfs:label "Is Calculation Element Of"@en ;
rdfs:range xebr:Total ;
.
xebr:isMonetaryPartOf
rdf:type owl:FunctionalProperty ;
rdf:type owl:ObjectProperty ;
rdfs:comment "This functions marks the \"monetary\" parts of a
presentation part of a report" ;
rdfs:domain xebr:Monetary ;
rdfs:label "Is Monetary Part Of"@en ;
rdfs:range xebr:Abstract ;
.
xebr:isPartOf
rdf:type owl:FunctionalProperty ;
rdf:type owl:ObjectProperty ;
rdfs:comment "This functions marks the part of xEBR document/Report
in which the items should be included" ;
rdfs:domain xebr:Abstract ;
rdfs:label "Is Part Of"@en ;
rdfs:range xebr:Role ;
.
5 Representing Multilingual Information
One of the major assets of XBRL, and even more xEBR, is the fact that con-
ceptual information is equipped with (multilingual) labels, representing the way
the concepts are linguistically realised in the corresponding national legislation,
with the expectation that companies will use a very similar terminology in their
reports. Just to name an example, the Belgian XBRL taxonomy comes with la-
bels in four languages (English and the three national languages: Dutch, French
and German). The xEBR initiative supports the specification of multilingual
equivalents within the context of the identified shared concepts across various
legislations.
While the XBRL/xEBR support of multilingualism is definitely a major
achievement, we see that this information can be expressed only in a somewhat
cumbersome and redundant manner in the used XML code. Also the so-called
“link base” organisation of the label data7 is not able to express terminological
generalisations and relations between the used terms. Additionally, we would
like to use a representation language that is compatible with the one we used for
the representation of the concepts, their types and their relations for encoding
the multilingual terms (or XBRL/xEBB labels). We opted therefore for (at least
in a first step) SKOS-XL. SKOS has been designed as an OWL and RDF(S)
compatible vocabulary for representing light-weight ontologies, like thesauri and
terminologies. SKOS-XL is an extension of SKOS that allows to handle the clas-
sical rdfs:label as an object, upgrading thus the content of some OWL annotation
properties to an autonomous element that can be manipulated from several on-
tologies8 . And it seems that this is exactly what we need for representing the
XBRL/xEBR (multilingual) labels, as we illustrate with a few examples below.
All labels of xEBR taxonomy are then introduced as an instance of the
skosxl:Label class. We can then explicitly describe relations between those la-
bels, so for example that “Crediti Verso Soci Versamenti Ancora Dovuti” can be
considered an Italian equivalent to “Subscribed capital unpaid [Presentation]”.
Both labels are then marked as prefLabel for the concept “Subscribed Capital
Unpaid Presentation”. It is compliant with the guidelines of SKOS-XL that a
concept has more than one object marked as prefLabel, if each object represents
a term in a distinct language. We foresee the use of skosxl:altLabel for termino-
logical variants one can observe, for instance, in the actual reports generated by
companies.
xebr:L_SubscribedCapitalUnpaidPresentation
rdf:type skosxl:Label ;
rdfs:comment "English PrefLabel for P.1.1
SubscribedCapitalUnpaidPresentationSubscribe"@en ;
rdfs:label "Subscribed Capital Unpaid Presentation"@en ;
skosxl:isTermTranslationOf xebr:L_CreditiVersoSociVersamentiAncoraDovuti ;
skosxl:literalForm
"Subscribed capital unpaid [Presentation]"@en ;
.
xebr:L_CreditiVersoSociVersamentiAncoraDovuti
rdf:type skosxl:Label ;
rdfs:comment "Italian Pref Label for P.1.1.
SubscribedCapitalUnpaidPresentation" ;
rdfs:label "Crediti Verso Soci Versamenti Ancora Dovuti"@it ;
skosxl:isTermTranslationOf xebr:L_SubscribedCapitalUnpaidPresentation ;
7
See https://www.xbrl.org/2003/xbrl-linkbase-2003-12-31.xsd for more de-
tails.
8
See https://www.w3.org/TR/skos-reference/skos-xl.html for more details.
skosxl:literalForm
"Crediti Verso Soci Versamenti Ancora Dovuti"@it ;
.
rdf:type owl:Class ;
xebr:isPartOf ;
xebr:version "V5" ;
rdfs:label "Subscribed capital unpaid [Presentation]"@en ;
rdfs:subClassOf xebr:Abstract ;
skosxl:prefLabel xebr:L_SubscribedCapitalUnpaidPresentation ;
.
Using SKOS-XL for modelling the (multilingual) labels, we are thus in the
position of establishing generalisations and marking explicitly relations between
the various labels within a legislation and the terminological variants of such
concepts as used in concrete reports and other data sources, as exemplified in
the next section.
6 Multilingual Short- and Long-Term Data Example
Shared semantic concepts between up-to-date information on, e.g. stock ex-
changes, and the XBRL reported information can provide a very powerful picture
on the finances of a corporation. Those shared concepts can be utilized in order
to align those two types of information. It has been shown before that structural
as well as value-based similarities can be utilized to align a company’s financial
data across languages [3]. Concentrating on various reports in various languages
for one company for a specific year allows for the additional use of a simple
heuristics in order to detect multilingual term correspondences: the financial po-
sitions associated with terms have the same values. While this heuristic alone is
too simplistic and would lead to faulty results, combined with a semantic simi-
larity measure applied to its labels it can be a powerful tool to align short-term
and long-term reporting information.
German Figure English Source
Kurzfristige Vermögenswerte 24.861 Short-term assets BASF
Kurzfristige Vermögen 24.861 Bundesanzeiger
Umlaufvermögen 24.861 Total Current Assets DAX
Table 1. Example of heuristic-based alignment of information on one specific company
across sources
As shown in Table 1, the figure-based aggregation of multilingual information
is capable of uncovering terminological variants in German and English. Based
on this heuristic, i.e., the identical figure shown for this one company’s posi-
tion, we learn that “Kurzfristige Vermögenswerte” is equivalent in meaning to
“Kurzfristige Vermögen” as well as “Umlaufvermögen”. Since two of the sources
provided in Table 1 are bilingual, we can learn two English translations of those
German terms, namely “Short-term assets” and “Total Current Assets”. While
similarity measures might perform well on the first two German terms, their ca-
pability of identifying a similarity to “Umlaufvermögen” might be limited. Here
the multilingual aspect fosters alignment, since most (not purely letter-based)
similarity measures provide higher values for the two English terms. Mediated by
the corresponding xEBR concepts – “xebr:CurrentAssetsTotal” – these German
and English term variants can then be automatically linked to other languages,
such as the xEBR Spanish label “activo corriente” and the associated value can
be compared to values reported in the long-term financial report. Linking the
xEBR label to the English DAX terms allows for enriching the ontology with
terminological variation that can be useful for detecting terms in actual com-
pany reports. However, the method for this linking should be elaborated on by,
for instance, using distributional semantic models.
7 Conclusion
Representing the xEBR taxonomy as an OWL ontology enables the use of se-
mantic technologies. However, this benefit can only be fully leveraged if the
implicit semantics of the taxonomy are explicated in the ontology, an approach
that we present in this paper. One aspect is also that we are able to improve
the modularity of the original taxonomy, as can be seen in the fact that mone-
tary elements have now their own class hierarchy. And we can mark explicitly
the relations between labels by the use of specialized properties. We also show
how this formalization of an XML-based format (or in case of xEBR, an Excel
based format) can foster the alignment of short-term reporting information to
long-term XBRL reporting by means of a brief example of stock exchange data.
One of the major benefits of this representation format is the ability to au-
tomatically check the consistency of data redundancy and generate financial
summaries - short- and long-term. Using xEBR as an ontology allows for the
comparison of companies across jurisdiction and across languages. If stakeholders
from different countries wish to obtain detailed information on their investment
in their native language, the proposed multilingual representation of financial
information can be very handy for human analysts.
Acknowledgement
We would like to thank the XBRL Europe Business Registers Working Group
(xEBR), and more specially Thomas Verdin, for their exciting work on the tax-
onomy and for providing us with the latest version of the Core Reference Taxon-
omy. We also thank Hans-Ulrich Krieger for his ground-breaking and inspiring
work on the ontologisation of XBRL. Our thank goes also to the anonymous
reviewers for their helpful comments on our submission.
References
[1] Bai, L., Koveos, P., Liu, M.: Applying an ontology-augmenting xbrl model to
accounting information system for business integration. Asia-Pacific Journal of
Accounting & Economics pp. 1–23 (2016)
[2] Bao, J., Rong, G., Li, X., Ding, L.: Representing financial reports on the semantic
web. In: International Workshop on Rules and Rule Markup Languages for the
Semantic Web. pp. 144–152. Springer (2010)
[3] Declerck, T., Gromann, D.: Extraction of multilingual term variants in the busi-
ness reporting domain. In: Proceedings of CHAT 2012: The 2nd Workshop on the
Creation; Harmonization and Application of Terminology Resources; Co-located
with TKE 2012; June 22; 2012; Madrid; Spain. pp. 41–46. No. 072, Linköping
University Electronic Press (2012)
[4] Declerck, T., Krieger, H.U.: Translating xbrl into description logic. an approach
using protege, sesame & owl. In: BIS. pp. 455–467 (2006)
[5] Fourny, G.: The XBRL Book – Simple, precise, technical. Ghis-
lain Fourny (1 2017), the first two chapters can be downloaded at
https://ghislainfourny.github.io/the-xbrl-book/download/the-xbrl-book.pdf
[6] Garcı́a, R., Gil, R.: Publishing xbrl as linked open data. In: CEUR Workshop
Proceedings. vol. 538. Citeseer (2009)
[7] Krahel, J.P.: On the formalization of accounting standards. Ph.D. thesis, Rutgers
University-Graduate School-Newark (2012)
[8] Liu, C., Luo, X.R., Wang, F.L.: An empirical investigation on the impact of xbrl
adoption on information asymmetry: Evidence from europe. Decision Support
Systems 93, 42–50 (2017)
[9] O’Riain, S.: Semantic Paths in Business Filings Analysis. Ph.D. thesis, National
University of Ireland, Galway (2012)
[10] O’Riain, S., Curry, E., Harth, A.: Xbrl and open data for global financial ecosys-
tems: A linked data approach. International Journal of Accounting Information
Systems 13(2), 141–162 (2012)
[11] Santos, I., Castro, E., Velasco, M.: Xbrl formula specification in the multidimen-
sional data model. Information Systems 57, 20–37 (2016)
[12] Spohr, D., Hollink, L., Cimiano, P.: A machine learning approach to multilingual
and cross-lingual ontology matching. In: International Semantic Web Conference.
pp. 665–680. Springer (2011)
[13] Thomas, S.M., Wu, X., Ma, Y., O’Riain, S.: Semantically assisted XBRL-
taxonomy alignment across languages. In: Buitelaar, P., Cimiano, P. (eds.) To-
wards the Multilingual Semantic Web, pp. 277–293. Springer (2014)
[14] Trojahn, C., Fu, B., Zamazal, O., Ritze, D.: State-of-the-art in multilingual and
cross-lingual ontology matching. In: Towards the Multilingual Semantic Web, pp.
119–135. Springer (2014)