Integration of a Domain Ontology in e-Science with a
Provenance Model for Semantic Provenance Generation in
the Scientific Images Analysis
Lucélia de Souza1,2, Maria Salete Marcon Gomes Vaz2,3
1
Departament of Computer Science – University of Western of Parana (UNICENTRO )
Rua Camargo Varela de Sá, 03, CEP 85040-080 – Guarapuava/PR – Brazil
2
Department of Informatics – Federal University of Parana (UFPR) Rua Cel. F. H. dos
Santos, 100, CEP 81.531-980 – Curitiba/PR - Brazil
3
Department of Informatics – State University of Ponta Grossa (UEPG) Av. Carlos
Cavalcanti, 4748, CEP 84.030-900 - Ponta Grossa/PR - Brazil
{lucelias,salete}@inf.ufpr.br
Abstract. This paper describes the integration of a domain ontology in e-Science with a
provenance model for semantic provenance generation in the scientific images analysis.
The domain ontology is related with images obtained from the CoRoT Telescope, where
the exoplanets search require detrend algorithms as preprocessing, improving the chance
to detect planetary transits. In order to retrieve standardized information regarding the
origin and facilitate the monitoring of information, the Proof Markup Language – PML
was chosen as provenance common model due to its characteristics of modularity, reuse,
interoperability and the possibility of justify how conclusions were obtained. As
contribution of this paper, the integration of the ontologies presented enables getting
information from the domain and justify the conclusions, through a standardized
provenance model, allowing logical inference and semantic interoperability.
1. Introduction
In the scientific images analysis, information of provenance provide the source of
processing, allowing share, reuse, reprocessing and do further analysis in data and
process. The semantic provenance [Sahoo et al 2008] is related with the Semantic Web
and can be obtained by means of ontologies [Borst 1997], which represent the
knowledge, structuring information in an organized manner and generating semantic in
the data.
In this work, a domain ontology was developed in the Ontology Web Language –
OWL2 called CorotDataAnalysisOntology (crtdao) [de Souza et al 2011] allowing to
extract domain information in the scientific images analysis. It relates with images
obtained from CoRoT Telescope1, which provides thousands of light curves in format
Flexible Image Transport System – FITS [Hanisch et al 2001]. In the analysis of these
images, the search of planets outside of Solar System (exoplanets) requires detrend
1
CoRoT Archive: http://idoc-corotn2-public.ias.u-psud.fr/
96
and/or filter algorithms as preprocessing for removing phenomena that may occur
suddenly, such as random jumps and/or trends, slow and gradual changes in certain
properties of the images, under whole range of the investigation. So, different detrend
and/or filter algorithms can be applied to treatment of these phenomena, improving the
chance to detect planetary transits (Figure1).
Figure1. CoRoT Data Analysis Process
However, domain ontology can be insufficient, semantically, for generation and
sharing of provenance information. It is necessary to make use of a common model for
the provenance generation as means to allow interoperability, reuse and extension of
ontologies [McGuinness et al 2007]. In this case, provenance models can be integrated
with domain ontologies because their use increases the understanding of users about how
answers were generated and also facilitates the acceptance of the results. Among the
provenance models existing, Provenir [Sahoo et al 2009], Open Provenance Model –
OPM [Moreau et al 2011] and Proof Markup Language – PML [McGuinness et al 2007]
stands out, being workflow-based systems. These models were analyzed for use in the
scientific images analysis, being chosen the PML due to its characteristics of modularity,
reuse, interoperability and mainly by allow us to justify how were obtained the
conclusions.
The objective of this paper is to present the integration of the domain ontology
with the PML Provenance Model, contributing to enrich the scientific images analysis
with semantic and standardization. This integration enables getting the information from
the domain and justifies the conclusions, by means of inference steps involving which the
inference engine, inference rule and/or source used to generate it.
This paper is structured as follows, besides this introductory section. The second
section describes about data provenance and workflows. The third section describes
about domain ontologies and provenance models and the next section presents their
integration. The fifth section brings the related works, followed of the conclusions and
the future works.
2. Data Provenance and Workflows
Provenance means origin or source. In the scientific images analysis, provenance
information proves the correctness of the resulting data, being regarded by Tan (2007) as
important as the result itself.
The provenance information can have granularity in the fine-grain and coarse-grain
forms [Tan 2007]. The first form involves the data derivation and storage in databases
how proposed by authors Tan (2007), Buneman et al (2007), Cheney et al (2009),
among others. There are two approaches, such as metadata annotation and non-
97
annotation approach, through queries and inverse functions used for data
transformations. The second form involves activities and processes used to perform tasks
of complex scientific data by mean of scientific workflows [Davidson and Freire 2008],
where can be human interactions during the execution of processes flow.
2.1 Semantic Provenance in the Scientific Images Analysis
This paper stands out by enriching the data analysis semantically, related to FITS images
that are available in the CoRoT Archive to exoplanets search. During the execution of
detrend and/or filter algorithms, provenance information can be stored in the FITS
images header.
The FITS Standard [Hanisch et al 2001] establishes rules for use of these images,
which differs of the traditional format of images, due to its basic structure formed by a
header containing metadata (data about data) such as SIMPLE, BITPIX, COMMENT,
HISTORY, among others, and a matrix used for storing binary data.
However, the FITS specification does not contemplate the addition of
provenance metadata, describing the use of HISTORY metadata to store steps executed.
This form of provenance generation is free text, not being machine readable, impeding its
use by software agents.
In scientific images analysis, provenance information records steps performed and
generating knowledge in order to avoid reprocessing and contribute to sharing, reuse and
analysis further. So, the metadata storing in images header or in the databases is
insufficient, semantically, to generate provenance. This information is useful for local
researchers, but not enough to share, reuse and reprocessing by scientific community.
There is a need for standardization of provenance metadata to be generated and stored,
just as it takes more detailed information to contemplate the real needs of researchers as
to the semantic knowledge about the data generation over time. Accordingly, the next
section describes the development of domain ontology in this environment.
2.2 Domain Ontology in e-Science
Ontology is defined as a formal and explicit specification of a shared conceptualization
[Borst 1997]. It is characterized as a mean of representing knowledge, structuring
information in an organized manner of a domain and generating semantics in the data.
The development process of the domain ontology proposed is based on Ontology
Development 101 [Noy and McGuinness 2001]. We started by identifying a set of
competency questions from domain that must be answered by the ontology, such as:
What are the statistical techniques (Linear, Polynomial, among others) used by
detrend algorithms?; The CoRoT Detrend Algorithm treats which systematic effects?;
What transit algorithm had the type of method Least-Squares?; among others. From
these questions, were identified classes, their relationships and the instances. Restrictions
are declared using axioms and/or rules, providing semantics and allowing inferences.
The Protégé 4.1 tool [Knublauch et al 2004] was used to the development of the
domain ontology and the generation of knowledge base. The used language is OWL 2.0,
98
recommend by World Wide Web Consortium - W3C and based on Descriptive Logic –
DL [Baader 2003]. Pellet 2.2 is used to verify its consistency.
An OWL ontology in abstract syntax contains annotations, axioms and facts.
However, the use only of axioms presents expressive limitations, mainly with the use of
properties such as composition of roles [Horrocks et al 2005].
The composition of roles as ‘isAlgorithmDetrendPolynomialOf’ shows an example.
If an algorithm is AlgorithmDetrend and the method type is Polynomial, then the algorithm
is an AlgoritmDetrend of the Polynomial type. The relationship between the composition
of the ‘isAlgorithmDetrendOf’ and ‘isMethodTypePolynomialOf’ properties and the
‘isAlgorithmDetrendPolynomialOf’ property is limited to the form PQ P, in order to
maintain decidability. The composition of two properties is a subproperty of one of the
composed properties, that is, the complex relationship between composed properties
cannot be captured. This is the case of ‘isAlgorithmDetrendPolynomialOf’ property that
cannot be captured because it is not one of ‘isAlgorithmDetrendOf’ not
‘isMethodTypePolynomialOf’.
So, the complex axiom ‘isAlgorithmDetrendOf’ ‘isMethodTypePolynomialOf’
‘isAlgorithmDetrendPolynomialOf’ presents the form RS T and TS R, because exists
cyclical dependences in the definition, violating the irreflexivity. This verification is
important in relation of the decidability, because such cyclical dependences can induce
undecidibility and the use in an ontology should be restricted. One way to address this
problem is extend OWL with a more powerful language to describe properties.
Horrocks et al (2005) extends axioms OWL DL to allow rule axioms (a Semantic
Web Rule Language - SWRL), in the form: axiom::=rule. In the human readable syntax,
a rule has the form antecedentconsequent, an implication between an antecedent
(body) and consequent (head).
Informally, a rule means “if the antecedent hold (is true), then consequent must
also hold”. The antecedent and consequent of a rule consist of zero or more atoms,
which can be of the form C(x), P(x,y), sameAs(x,y) or differentFrom(x,y), where C is an
OWL DL description, P is an OWL property, x and y are either variables, individuals or
data values. Multiple atoms in an antecedent are treated as conjunction and multiple
atoms in a consequent are treated as separate consequences [Horrocks et al 2005].
The Protégé 4.1 Tool allows working with rules from View Rules. Pellet
supports reasoning with SWRL rules, which interprets SWRL using the DL-Safe Rules
notion, where rules will be applied only to named individuals in the ontology.
2.3 Analysis of the domain ontology as to semantic integration
The domain ontology was evaluated by domain experts as the terms used as well by
ontologists. Also was formalized in an extension of the DL called SROIQ [Horrocks et al
2006], which presents characteristics of the expressiveness, decidability and robust
computational properties, being an extension more expressive than the Attributive
Language, the most basic family of DL.
The formalization allow us to specify the ontology independently of the domain,
contributing to verification and validation of axioms assertional, terminological and role
99
inclusion used, well as allows to infer knowledge. With the formalization in SROIQ DL,
under OWL 2, recommended in 2009 by the W3C, also is possible to verify the
consistency of the knowledge base. In this way, it is feasible to enrich the scientific
images analysis with semantic and standardization.
The domain ontology proposed in [de Souza 2011] presents as main classes:
DataSet, Methods, Technique, AlgorithmBase, PeriodicSignalShape, MethodType,
Algorithm, Software, Metadata, Person, Run, Telescope, Language and
SistematicEffectType. Header class was created to relate header specific metadata of
FITS images and the Database class, related with the storage location. The Language
class was specified in ProgramationLanguage. However, aiming semantic integration in
e-Science, a domain ontology developed was evaluated in relation to existing ontologies
(Figure2).
The VSTO ontology2 stands out as an ontology open-source, extensible and
reusable in the area of solar-terrestrial physics, which supports interdisciplinary projects
of virtual data collections. This ontology was analyzed, and made the following
adjustments in the domain ontology: i. Telescope class was inserted as a subclass of
Instrument and were also imported from VSTO the following classes: DataProduct
related with FITS images, which were previously represented as DataSet;
vsto:InstrumentOperationMode related with information about the operation mode of
the instrument; vsto:DateTimeInterval, being intervals for date and time and
vsto:Parameter, including the following parameters: ErrorParameter, Noise, Period,
SignalToNoiseRatio, TimeDependentParameter and StatisticalMeasure.
The Semantic Web Earth and Environmental Terminology - SWEET Ontology3
has widespread acceptance in e-Science. However, this ontology extends more in width
than depth in certain areas. Thus, for purposes of interoperability and reuse, the
crtdao:MethodType class was replaced by import of the sweet:Process class. It’s
because the objective of this work is to deepen concepts to generate semantic
provenance as the statistical methods used in the analysis of FITS images.
3. Domain Ontologies and Provenance Models
Domain ontologies should be built based on Foundation Ontology, such as SUMO4,
DOLCE5, UFO6, among others, because they are theoretically well-founded, becoming
the category systems independent of domain, describing the general concepts and
improving the quality of conceptual model [Guizzardi 2005]. They are characterized by
being highly reusable because it shapes basic and general concepts, as well as relations.
However, the well-founded ontologies are generic about many areas.
So, due to the need for representing provenance information, provenance models
stands out because are ontologically well-founded representation models, adding
concepts and relationships provenance-aware, allowing the adoption of a common
provenance terminology [McGuinness et al 2007]. These models are presented follow.
2
Virtual Solar-Territorial Observatory: http://escience.rpi.edu/ontology/vsto/2/0/vsto.owl
3
Semantic Web Earth and Environmental Terminology: http://sweet.jpl.nasa.gov/
4
Suggested Upper Merged Ontology: http://www.ontologyportal.org/
5
Descriptive Ontology for Linguistic and Cognitive Engineering http://www.loa.istc.cnr.it/DOLCE.html
6
Unified Foundational Ontology: http://code.google.com/p/ufo-nemo-project/
100
3.1 Open Provenance Model - OPM
It is an abstract model developed from Provenance Challenge Series to explain how
artifacts were derived, based on workflows. It is independent of technology for
interoperability purposes. Uses a graph based on a syntactic rules set and topological
constraints. It presents as concepts Agent, denoting people; Process, denoting actions or
executions of process; and Artifacts, denoting the entity produced or manipulated. This
data model has applicability mainly in biologic area.
The modularity of this data model involves OPM Specification, OPMV
Vocabulary, OPMO Ontology and XML Schema. The focus is on provenance in
workflows, defining a small set of key concepts to general entities and relationships
(wasGeneratedBy - WGB and WasControledBy - WCB) in workflows. On the
downside, the OWL Profile is still evolving to adapt the OPM Specification.
3.2 Provenir
This ontology presents as main concepts Agent, Process and Data. Data_Collection and
Parameters spatial, domain and temporal are subclass of the Data. It is constituted by
eight classes and eleven properties, including the Relation Ontology.
It presents as characteristics a common model to represent provenance, being
expressive as the concepts and relationships modeled named well-defined, can be
extended to modeling of complex provenance information and domain-specific, enabling
analysis in SWRL and W3C Rule Interchange Format - RIF. This ontology has
applicability in biomedical and oceanography areas in real projects of the e-Science.
3.3 Provenance Markup Language - PML
PML is based on Proof Theory and constitutes a common model for represent and share
explanations generated by various intelligent systems such as answers systems of hybrid
web questions, analytical text, theorem provers, among others. It describes the
justifications as a sequence of information manipulations steps used to generate a
response. This sequence is referred as a proof.
Due to modularity, it is possible to use modules individually for Provenance
(PML-P), Justification (PML-J) or hold Trust (PML-T) in the data7. The PML-T
supports annotation of complex trust relations in provenance concepts and justifications.
The primitive concepts and relations are specified in OWL, facilitating reuse and
extension. The modules PML-P and PML-J are described in the following.
3.3.1 Provenance Ontology - PML-P
PML provides a vocabulary for justification of metadata whose focus is on
representational primitives used to describe properties of 'things' identified as
information, language and resources, such as organization, person, agent and services.
These primitives are extensible, used to annotate the source of information, as to
represent sources used and who encoded the information. PML-P presents the following
concepts.
7
URL: http://inference-web.org/2007/primer
101
An instance of IdentifiedThing refers to a real world entity and its properties
note the properties of entities such as name, description, date and time of creation and
ownership. PML-P also includes Information, Source and SourceUsage, Language and
InferenceRule subclasses.
The Information subclass supports references to information on various levels of
granularity and structure, such as a formula in a logical language, a fragment of natural
language or a dataset. The Source is extensible and refers to a container of information,
such as a Document, an Agent, among others. SourceUsage is used to associate
Information and Source, declaring information from a Source at certain time. Language
represents the language in that the conclusion is represented. InferenceRule aims to
encode various types of computation steps.
3.3.2 Justification Ontology - PML-J
This module requires concepts to represent conclusions, zero or more sets of antecedents
of the conclusion and the steps used to manipulate information to get conclusions from
the set of antecedents and so on recursively. The vocabulary for explanations of data
focuses on representational primitives used to explain dependencies between 'things',
including constructors to represent how conclusions are derived. It presents the NodeSet
and InferenceStep concepts.
The NodeSet represents a conclusion and a set of alternative steps, each of which
may provide an alternative justification for a conclusion. This term captures the concept
of a set of nodes in steps from one or more proof trees deriving the same conclusion.
An InferenceStep represents a justification for the conclusion of the respective
NodeSet. It refers to a logical step of inference, an information extraction step, any step
in the process of computing, or an assertion of a fact or an assumption. It can also be a
complex process as web service or application. An InferenceStep represents the details
such as the InferenceEngine, InferenceRule, and the set of antecedents NodeSets of one
justification for the conclusion of the corresponding NodeSet.
4. Integration of the Domain Ontology with the Provenance Model PML
In this work, we choose to make use of the PML-P and PML-J modules of PML model,
mainly because allows us represent and explain how the conclusions were obtained by
informing which the inference engine, the rules and the source of information used, as
well as due to modular design.
The integration (Figure2) is done using multiple inheritance of the classes as in
Zednik el al (2009), where an individual is defined as a type from the provenance model
and at least one type from the domain ontology, e.g. the CoRoT instance is defined as
belonging to classes crtdao:Telescope and the extension pmlp:Telescope, being an
subclass of pmlp:Agent of the pmpl:Source class. So, an instance of crtdao:Telescope
becomes the source used to justify a conclusion from a NodeSet. The same classes in
crtdao e pmlp are treated as equivalent classes.
From a Question is created the respective Query, which is linked to a NodeSet,
where is stated a conclusion (Information) through the hasConclusion property.
NodeSet may have none, one or more InferenceSteps stating which InferenceRule,
InferenceEngine and/or Source were used, beyond of a list of antecedents NodeSets.
102
Figure2. Integrated Ontologies visualized in OntoGraf Plugin of Protégé 4.1 Tool
Given the Question ‘What is the source of a given dataset?’ is specified the Query in a
given language binds to a NodeSet and declares as conclusion the respective source. As
inference step, it is possible to declare the Source of the using the property hasSource.
So, it is possible to justify the source of information in the standardized way.
McGuinness et al (2007) identifies four types of justifications for a given
conclusion, exemplified below in XML format and Protégé 4.1 Tool:
i. The conclusion is an unproven conclusion or goal. No justification is available and
none InferenceStep is associated with the NodeSet. For the Question What is the
technique of the Photometric Detrend Algorithm? and the Query in the Manchester
OWL DL Query syntax (Figure3), the conclusion is given by NodeSet respective using
the properties pmlp:hasLanguage related with the language of the conclusion and
pmlp:hasRawString related with the content of information as a string.
(English)>
(Photometric_Detrending_Algorithm
hasTechnique value Photometric)
Figure3. Justifying a unproven Conclusion without InferenceStep
ii. The conclusion is an assumption. The conclusion is directly assumed by an agent as a
true statement. The Question What Methods the SARS algorithm belongs to? is justified
by inference in the NodeSet respective that includes the information pmlp:hasRawString
and pmlp:hasLanguage. As a consequence of the InferenceStep is declared assumption
as InferenceRule and Pellet as InferenceEngine (Figure4).
103
(SARS_Algorithm hasMethods value
Data_Analysis)
(English)>
Pellet
assumption
Figure4. Justifying a Conclusion using InferenceStep
iii. The conclusion is a direct assertion. It can be declared by Inference Engine directly
without using any antecedent information (Figure5). For the Question What is the
publication of Corot Detrend Algorithm?, the NodeSet respective declare the
information using pmlp:hasRawString and pmlp:hasLanguage properties. As the
consequence, the InferenceStep informs direct_assertion using pmlj:hasInferenceRule
and the pmlp:hasDocument property informs the Source. Also it is possible to declare
other details about the publication how number of pages and URL.
(Corot Detrend Algorithm hasPublication value An algorithm for
correction CoRoT raw light curves)
(English)>
direct_assertion
An algorithm for correction…
1
8
…
Figure5. Justifying a Conclusion using Direct Assertion
iv. The conclusion is derived from a list of antecedents by applying a certain
computation. This representation to encode many types of computation steps. The
Question Which is the function type of the Corot Detrend Algorithm? (Figure6) shows
that the conclusion is derived from first NodeSet or from rest NodeSet.
(English)>
Corot_Detrend_Algorithm
hasFunction polynomial
nsCorotDetrendAlgoritmPublication
0
nsDetrendpolynomial
1
…
Figure6. Justifying a Conclusion derived from AntecedentList
104
5. Related work
Zednik et al (2009) present how the semantic provenance is reconstructed to data
products in coronal physics area. This work provides a foundation for scientific
workflow provenance applications, describing the use of semantic web technologies to
encode provenance and domain information and demonstrating how both can be used
together to satisfy complex use case. The data model use OWL ontologies independent.
The Solar-Terrestrial Ontology – VSTO is used as a core domain model in e-Science,
modeling data products, instruments and parameters. The provenance model uses the
Inference Web and the Framework PML, chosen because of its capacibilities of represent
conclusions, justificatives and explanations. The integration of provenance and domain
models is done by means of multiple-inheritance from individuals´ declarations of the
ontologies. The search results can be seen by Inference Web browser or by Probe-It!,
enabling scientists to better understand imperfections and processing consequences upon
e-Science data images.
Malaverri et al (2012) presents an approach of provenance to ensure the quality
of geospatial data, combining features provided by the OPM and FGDC geographic
metadata standards. It presents a case study in agriculture area, considering the
trustworthiness of source, is that, the degree of confidence of who created/made
available the data and temporality dimensions including valid and transaction time, e.g.
‘when’ related to data quality. Despite the proposal model be based on OPM model, is
added own characteristics taking into account the geospatial domain and assessment of
data quality. As future works, techniques to compute and assess the trustworthiness of
data will be investigated.
Salayandia et al (2012) propose a framework to support the creation of
ontologies for management of scientific data, specifying an abstraction in the form of a
top-level ontology codified in OWL-DL, including general concepts that can be
specialized to describe the capture and transformation of data. The Ontology Driven
Workflow (WDO.owl) is proposed and presents three basic concepts: Date, things that
can be used directly or indirectly as evidence, e.g. the output of a sensor; Method, things
that can be used to transform the data, e.g. visualization of software; and Container,
things that can be used as acquires or placeholders of the data, e.g. a database. WDO is
specified in Description Logic and the knowledge representation system is divided into
Tbox terminology and Abox, including assertions as the individuals in relation to the
Tbox. WDO is aligned with PML, where the concepts Date, Container and Method are
included respectively by PML concepts: Information, Source and Inference Rule. The
formalism that aligns the WDO and PML Ontologies is also specified using DL,
including subsumption equations rather than equalities due to concepts related with the
provenance are more general than the concepts of the WDO Ontology. It is because data
can be transformed by systematic processes, where the framework can be used to
document the process.
This paper stands out by enrich with semantic and standardization the phases of
the detrending and exoplanets search, providing information about the semantic
provenance of data and statistical methods used in the correction and analysis of FITS
images, contributing for adding semantic knowledge in experiments of e-Science and
take advantage of the features provided by PML.
105
6. Conclusions
It is presented in this paper that is possible to generate semantic provenance in scientific
images analysis. The environment involves FITS images from CoRoT Archive and the
integration of data models related to domain ontology and the provenance model. This
integration allow us make use of a common model and standardized for generating
provenance, contributing for semantic interoperability and allowing us to justify how
conclusions were obtained in the knowledge base.
Due to need for representing provenance information, provenance models are
ontologically well-founded, adding concepts and relationships provenance-aware,
allowing the adoption of a common provenance terminology. In this work, we choose to
use the PML model by allowing us to represent and explain how the conclusions were
obtained providing the inference engine, the inference rules and the source used, as well
as due to its modularity.
The semantic provenance information obtained will be persisted in databases, and
integrated in a web framework, facilitating the information retrieval processes, where
queries of provenance can be performed, allowing further analysis and contributing to
enrich semantically the development of scientific experiments. Despite the scope of this
work, results can be expanded to fields of e-Science where the scientific images analysis
requires preprocessing, adding semantic knowledge and allowing interoperability.
Acknowledgments
We acknowledge the support of the Astronomical Observatory AstroUEPG.
References
Baader, F. and Calvanese, D. and McGuinness, D. L. and Nardi, D. and Patel-Schneider,
P. F. (2003). The Description Logic Handbook: Theory, Implementation, and
Applications. Cambridge University Press, New York, USA.
Borst, W. N. (1997). Construction of Engineering Ontologies for Knowledge Sharing
and Reuse. Doctoral Thesis. University of Tweenty, Centre for Telematica and
Information Technology, Enschede, The Netherlands, (227 pp.).
Buneman, P. and Khanna, S. and Tan, W. C. (2007). Why and Where: A
Characterization of Data Provenance. In ICDT (pp. 316–330).
Cheney, J. and Chiticariu, L. and Tan, W. C. (2009). Provenance in Databases: Why,
How, and Where. Foundations and Trends in Databases (Vol. 1, N. 4, pp. 379–474).
Hanover, MA, USA: Now Publishers Inc.
Davidson, S. B. and Freire, J. (2008). Provenance and Scientific Workflows: Challenges
and Opportunities. In Proceedings of ACM SIGMOD (pp. 1345–1350).
De Souza, L. and Vaz, M. S. M. G. and Emílio, M. and da Rocha, J. C. F. and Boufleur,
R. (2011). Data Analysis Provenance: Use Case for Exoplanet Search in CoRoT
Database. Presented In: ADASS XXI Conference Series. In Press: ASP Conf. Ser.
Vol. TBD. San Francisco. 2012.
106
Guizzardi, G. (2005) Ontological Foundations for Structural Conceptual Models. PhD
Thesis (CUM LAUDE), University of Twente, The Netherlands. Published as the
book Ontological Foundations for Structural Conceptual Models, Telematica Instituut
Fundamental Research Series No. 15.
Hanisch, R. J. and Farris, A. and Greisen, E. W. and Pence, W. D. and Schlesinger, B.
M. and Teuben, P. J. and Thompson, R. W. and Warnock III, A. (2001). Definition of
the Flexible Image Transport System (FITS). A&A (Vol. 376, pp. 359–380).
Horrocks, I. and Kutz, O. and Sattler, U. (2006). The Even More Irresistible SROIQ.
Proc. of the 10th Int. Conf. on Principles of Knowledge Representation and
Reasoning (KR 2006), (pp. 57–67), AAAI Press.
Horrocks, I. and Patel-Schneider, P. F. and Bechhofer, S. (2005) OWL Rules: A
Proposal and Prototype Implementation. Journal of Web Semantics, V. 3, N. 1.
Knublauch, H. and Fergerson, R. W. and Noy, N. F. and Musen, M. A. (2004). The
Protégé OWL Plugin: An Open Development Environment for Semantic Web
Applications. 3rd ISWC 2004, Hiroshima, Japan. (pp. 229–243).
Malaverri, J. E. G. and Medeiros, C. B. and Lamparelli, R. C. (2012) A provenance
Approach to Assess Quality of Geoespatial Data. Symposium on Applied Computing.
McGuinness, D. L. and Ding, L. and da Silva, P. P. and Chang, C. (2007). PML 2: A
Modular Explanation Interlingua. In Proc. of the AAAI 2007 Workshop on
Explanation-aware Computing (pp. 22–23).
Moreau, L. and Clifford, B. and Freire, J. and Futrelle, J. and Gil, Y. and Groth, P. and
Kwasnikowska, N. and Miles, S. and Missier, P. and Myers, J. and Plale, B. and
Simmhan, Y. and Stephan, E. and den Bussche, J. V. (2011). The Open Provenance
Model Core Specification (v1.1). Future Generation Computer Systems (Vol. 27, N.
6, pp. 743–756). Amsterdam, The Netherlands: Elsevier Science Publishers B. V.
Noy, N. F. and McGuinness, D. L. (2001). Ontology Development 101: A Guide to
Creating Your First Ontology. Stanford Knowledge Systems Laboratory Tech. Rep.
KSL-01-05 and Stanford Medical Informatics Tech. Rep. SMI-2001-0880 (25 pp.).
Sahoo, S. S. and Sheth, A. and Henson, C. (2008). Semantic Provenance for eScience.
IEEE Computer Society, pp. 46-54.
Sahoo, S. S. and Weatherly, D. B. and Mutharaju, R. and Anantharam, P. and Sheth, A.
and Tarleton, R. L. (2009). Ontology-Driven Provenance Management in eScience:
An Application in Parasite Research. Proc. of the CoopIS, DOA, IS, and ODBASE
2009 on On the Move to Meaningful Internet Systems: Part II. OTM ’09. (pp. 992–
1009). Berlin, Heidelberg: Springer-Verlag.
Salayandia, L. and Pinheiro, P. and Gates, A. Q. (2012). A Framework to Create
Ontologies for Scientific Data Management. University of Texas at El Paso.
Tan, W. C. (2007). Provenance in Databases: Past, Current, and Future. IEEE Data Eng.
Bull. (Vol. 30, N. 4, pp. 3-12).
Zednik, S. and Fox, P. and McGuinness, D. L. and da Silva, P. P. and Chang, C. (2009).
Semantic Provenance for Science Data Products: Application to Image Data Processing.
First International Workshop on the role of Semantic Web in Provenance Management.
107