<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>A GUI for visualising and manipulating multiple ontology alignments</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Bernardo Severo</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Cassia Trojahn</string-name>
          <xref ref-type="aff" rid="aff1">1</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Renata Vieira</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Pontif cia Universidade Catolica do Rio Grande do Sul</institution>
          ,
          <addr-line>Porto Alegre</addr-line>
          ,
          <country country="BR">Brazil</country>
        </aff>
        <aff id="aff1">
          <label>1</label>
          <institution>Universite Toulouse 2 &amp; IRIT</institution>
          ,
          <addr-line>Toulouse</addr-line>
          ,
          <country country="FR">France</country>
        </aff>
      </contrib-group>
      <abstract>
        <p>This paper presents a Web-based environment for visualising and manipulating multiple ontology alignments at schema-level. While most solutions are limited to the visualisation of single alignments and are still provided as part of speci c standalone matching systems, we propose an open environment that is not bound to any speci c system. Within this graphical environment, users can manually create, suppress and edit correspondences and apply a set of operations on alignments ( ltering, merge, di erence, etc.). Evaluating multiple alignments, against a reference one, can also be carried out using classical evaluation metrics. In this demo, all these tasks will be demonstrated.</p>
      </abstract>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>Introduction</title>
      <p>
        Ontology matching has become an essential task in the management of the
semantic heterogeneity problem in open environments. Diverse matching
approaches have been proposed in the literature [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ] and systematic evaluation of
them has been carried out over the last ten years [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ]. Despite the progress in the
eld, visualisation of ontology alignments is still an open question. Nowadays,
most solutions for (graphical) alignment visualisation are limited to the
visualisation of single alignments and are still provided as part of speci c standalone
matching systems [
        <xref ref-type="bibr" rid="ref1 ref8">1, 8</xref>
        ]. Few systems expose their functionalities as Web-based
user interfaces [
        <xref ref-type="bibr" rid="ref11 ref7">7, 11</xref>
        ]. Alternative solutions include Protege plugins [
        <xref ref-type="bibr" rid="ref9">9</xref>
        ] or
Webbased interfaces for creating, manipulating and storing alignments (with limited
support for graphical visualisation) [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ].
      </p>
      <p>
        In this paper, we present an extended version1 of VOAR (Visual Ontology
Alignment EnviRonment) [
        <xref ref-type="bibr" rid="ref10">10</xref>
        ] that implements the graphical visualisation of
multiple alignments at schema-level. VOAR provides a Web-based environment
that is not bound to any speci c system and that o ers a GUI for assisting users
in the tasks of (multiple) alignment visualisation, manipulation, and evaluation.
With this new feature, users can better compare and analyse alignments (i.e.,
parts of the ontologies which are covered for most alignments and those which
are not, consensus between alignments, etc.). The main modules of VOAR are
presented in the following.
      </p>
      <sec id="sec-1-1">
        <title>1 http://voar.inf.pucrs.br/</title>
      </sec>
    </sec>
    <sec id="sec-2">
      <title>VOAR : Visual Ontology Alignment Environment</title>
      <p>
        The matching process takes two ontologies o and o0 as input and generates,
as output, a set of correspondences (i.e., an alignment) between them. Each
correspondence expresses the relation r (e.g., , w, etc.) between the entities e
and e0, with a con dence in the fact that the relation holds, typically in the [
        <xref ref-type="bibr" rid="ref1">1,0</xref>
        ]
range. In VOAR, alignments can either be created from scratch (by informing
the URIs of the two ontologies to be aligned) or loaded from an external le (in a
well-known RDF format2), which describes a set of metadata, including the URIs
of the aligned ontologies. The generation of alignments by matching systems is
out of the scope of this demonstration. VOAR does not store any alignment and
works on the notion of a `working alignment', an in-memory alignment in which
the results of the operations are reported (for instance, multiple alignments can
be merged together and the result of this operation is stored on the working
alignment). In the current version of VOAR, ontologies have to be in RDFS or
OWL and alignments are limited to correspondences involving concepts.
      </p>
      <p>
        One important aspect in alignment visualisation refers to the visualisation
of the ontologies themselves. VOAR adopts a commonly used technique that
consists in presenting the ontologies as indented trees, where indentation
represents the hierarchical relations between concepts (concepts and their
subconcepts can be expanded and retracted, in order to adjust the detail of information
shown). Although this kind of technique has advantages and disadvantages [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ],
it is known to be familiar to most users and o ers a clear view of the hierarchy3.
The correspondences are then shown as lines between the two trees.
      </p>
      <p>VOAR has been developed on the top of well-know and largely used APIs in
the eld of ontology matching, such as the Alignment API4 and the OWL API5,
and it is built with JavaServer Faces (JSF) technology, compatible with most
current browsers. Alignments in VOAR can be manipulated through four main
modules, which are brie y presented in the following.</p>
      <p>Correspondence edition This module allows for manipulating the
correspondences from one single working alignment. They can be added, suppressed or
edited (e.g, the kind of relation or con dence can be modi ed). The user can
select speci c ontology entities for visualising the correspondences involving them.
A new correspondence can be created by selecting entities in both ontologies.
In each tree, users can also lter out the entities whose naming and annotations
(labels, comments) correspond to a search criteria, what is especially useful when
dealing with large ontologies.</p>
      <p>Alignment manipulation VOAR provides di erent ways for manipulating a
single alignment as an `object', which can be combined together with other
`objects'. The current set of available operations are: union (resulting in the merge</p>
      <sec id="sec-2-1">
        <title>2 http://alignapi.gforge.inria.fr/format.html</title>
        <p>3 Here, we focus on the visualisation of concept hierarchies and alternative ways for
visualising whole ontologies are being considered.</p>
      </sec>
      <sec id="sec-2-2">
        <title>4 http://alignapi.gforge.inria.fr</title>
      </sec>
      <sec id="sec-2-3">
        <title>5 http://owlapi.sourceforge.net/</title>
        <p>A GUI for visualising and manipulating multiple ontology alignments
of alignments), intersection (where only correspondences occurring in all input
alignments are kept), di erence (complement, where a subset of correspondences
are removed from a given alignment). Auxiliary operations involve the possibility
to trim correspondences under a given threshold and invert the direction of the
alignment. These operations are available in the Alignment API and Alignment
Server for a pair of alignments, in their native versions, and we have reused them
here through the Alignment API. In VOAR, users can select multiple alignments
at once and apply those operations on the current working alignment.
Alignment evaluation Users can compare multiple alignments with respect to a
reference alignment. Measures of precision, recall, and F-measure can be
computed for each alignment. The implementations available in the Alignment API
have been used for generating these evaluation metrics. VOAR allows the
visualisation of the results in a tabular view and displays the status of each
correspondence (true negative, false negative, true positive and true negative) through
icons representing the entities involved in the given correspondence.</p>
        <p>Alignment visualisation From a set of previously loaded or manually created
alignments, this module allows for visualising multiple alignments together. The
set of correspondences of a given alignment is represented by a di erent line
colour. The user is able to con gure the properties of the correspondences to be
shown (e.g., kind of relation and con dence threshold) and apply lters on those
properties in order to visualise a subset of correspondences. Figure 1 presents the
screenshot of the visualisation of three alignments generated by OAEI systems
(Aroma, LogMap and RiMOM) on the real Benchmark test case 3046.</p>
      </sec>
    </sec>
    <sec id="sec-3">
      <title>Conclusions and Future Work</title>
      <p>
        This paper has presented an extended version of VOAR, a web-based
environment for visualising and manipulating ontology alignments at schema level. The
major contribution of this new version is the possibility of visualising multiple
alignments together. It allows for better comparing and analysing alignments
(i.e., parts of the ontologies which are covered for most alignments and those
which are not, consensus between alignments, etc.). As future work, we plan to
take into account the visualisation of properties and entities at instance-level.
We intend as well to provide an evaluation on its usability. Finally, we plan to
deal with the scalability of VOAR in terms of the visualisation of large
ontologies and to evaluate how it satis es a set of requirements [
        <xref ref-type="bibr" rid="ref6">6</xref>
        ] for user support in
large-scale ontology alignment task.
      </p>
      <sec id="sec-3-1">
        <title>6 http://oaei.ontologymatching.org/tests/</title>
      </sec>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1.
          <string-name>
            <given-names>I.</given-names>
            <surname>Cruz</surname>
          </string-name>
          ,
          <string-name>
            <given-names>F.</given-names>
            <surname>Antonelli</surname>
          </string-name>
          , and
          <string-name>
            <given-names>C.</given-names>
            <surname>Stroe</surname>
          </string-name>
          .
          <article-title>AgreementMaker: e cient matching for large real-world schemas and ontologies</article-title>
          .
          <source>In Proc. International Conference on Very Large Databases</source>
          , pages
          <volume>1586</volume>
          {
          <fpage>1589</fpage>
          ,
          <year>2009</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          2.
          <string-name>
            <given-names>J.</given-names>
            <surname>David</surname>
          </string-name>
          ,
          <string-name>
            <given-names>J.</given-names>
            <surname>Euzenat</surname>
          </string-name>
          , F. Schar e, and
          <string-name>
            <given-names>C.</given-names>
            <surname>Trojahn</surname>
          </string-name>
          .
          <source>The Alignment API 4.0. Semantic Web</source>
          ,
          <volume>2</volume>
          (
          <issue>1</issue>
          ):3{
          <fpage>10</fpage>
          ,
          <year>2011</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          3.
          <string-name>
            <given-names>J.</given-names>
            <surname>Euzenat</surname>
          </string-name>
          ,
          <string-name>
            <given-names>C.</given-names>
            <surname>Meilicke</surname>
          </string-name>
          ,
          <string-name>
            <given-names>H.</given-names>
            <surname>Stuckenschmidt</surname>
          </string-name>
          ,
          <string-name>
            <given-names>P.</given-names>
            <surname>Shvaiko</surname>
          </string-name>
          , and
          <string-name>
            <given-names>C.</given-names>
            <surname>Trojahn</surname>
          </string-name>
          .
          <article-title>Ontology alignment evaluation initiative: Six years of experience</article-title>
          .
          <source>Journal of Data Semantics</source>
          ,
          <volume>15</volume>
          :
          <fpage>158</fpage>
          {
          <fpage>192</fpage>
          ,
          <year>2011</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          4.
          <string-name>
            <given-names>J.</given-names>
            <surname>Euzenat</surname>
          </string-name>
          and
          <string-name>
            <given-names>P.</given-names>
            <surname>Shvaiko</surname>
          </string-name>
          . Ontology matching. Springer, Heidelberg (DE),
          <year>2007</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          5.
          <string-name>
            <given-names>B.</given-names>
            <surname>Fu</surname>
          </string-name>
          ,
          <string-name>
            <given-names>N.</given-names>
            <surname>Noy</surname>
          </string-name>
          , and
          <string-name>
            <given-names>M.</given-names>
            <surname>Storey</surname>
          </string-name>
          .
          <article-title>Indented tree or graph? A usability study of ontology visualization techniques in the context of class mapping evaluation</article-title>
          .
          <source>In Proc. 12th International Semantic Web Conference</source>
          , pages
          <volume>117</volume>
          {
          <fpage>134</fpage>
          ,
          <year>2013</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          6.
          <string-name>
            <given-names>V.</given-names>
            <surname>Ivanova</surname>
          </string-name>
          ,
          <string-name>
            <given-names>P.</given-names>
            <surname>Lambrix</surname>
          </string-name>
          , and
          <string-name>
            <given-names>J.</given-names>
            <surname>Aberg</surname>
          </string-name>
          .
          <article-title>Requirements for and evaluation of user support for large-scale ontology alignment</article-title>
          .
          <source>In Proc. 12th European Semantic Web Conference</source>
          , pages
          <volume>3</volume>
          {
          <fpage>20</fpage>
          .
          <year>2015</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          7.
          <string-name>
            <given-names>E.</given-names>
            <surname>Jimenez-Ruiz</surname>
          </string-name>
          and
          <string-name>
            <given-names>B. C.</given-names>
            <surname>Grau</surname>
          </string-name>
          . LogMap:
          <article-title>Logic-based and scalable ontology matching</article-title>
          .
          <source>In Proc. 10th International Semantic Web Conference</source>
          , pages
          <volume>273</volume>
          {
          <fpage>288</fpage>
          ,
          <year>2011</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          8.
          <string-name>
            <given-names>D. H.</given-names>
            <surname>Ngo</surname>
          </string-name>
          and
          <string-name>
            <given-names>Z.</given-names>
            <surname>Bellahsene</surname>
          </string-name>
          . YAM+
          <article-title>+ : (not) Yet Another Matcher for Ontology Matching Task</article-title>
          . In 28e Journees Bases de Donnees Avancees, France,
          <year>2012</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          9.
          <string-name>
            <given-names>N.</given-names>
            <surname>Noy</surname>
          </string-name>
          and
          <string-name>
            <given-names>M.</given-names>
            <surname>Musen</surname>
          </string-name>
          . PROMPT:
          <article-title>Algorithm and tool for automated ontology merging and alignment</article-title>
          .
          <source>In Proc. 17th National Conference on Arti cial Intelligence</source>
          , pages
          <fpage>450</fpage>
          {
          <fpage>455</fpage>
          ,
          <year>2000</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          10.
          <string-name>
            <given-names>B.</given-names>
            <surname>Severo</surname>
          </string-name>
          ,
          <string-name>
            <given-names>C.</given-names>
            <surname>Trojahn</surname>
          </string-name>
          , and
          <string-name>
            <given-names>R.</given-names>
            <surname>Vieira</surname>
          </string-name>
          . VOAR:
          <article-title>A visual and integrated ontology alignment environment</article-title>
          .
          <source>In Proc. 9th International Conference on Language Resources and Evaluation</source>
          , pages
          <volume>3671</volume>
          {
          <fpage>3677</fpage>
          ,
          <year>2014</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          11.
          <string-name>
            <surname>J. Volz</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          <string-name>
            <surname>Bizer</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          <string-name>
            <surname>Gaedke</surname>
            , and
            <given-names>G.</given-names>
          </string-name>
          <string-name>
            <surname>Kobilarov. SILK -</surname>
          </string-name>
          <article-title>a link discovery framework for the web of data</article-title>
          .
          <source>In Proc. WWW Workshop on Linked Data on the Web</source>
          ,
          <year>2009</year>
          .
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>