9 results on '"Porn, Alex Mateus"'
Search Results
2. Teste de ontologias OWL : uma contribuição utilizando o teste de mutação
- Author
-
Porn, Alex Mateus, 1985, Peres, Leticia Mara, 1972, and Universidade Federal do Paraná. Setor de Ciências Exatas. Programa de Pós-Graduação em Informática
- Subjects
Ciência da Computação ,Ontologia ,Modelagem de dados - Abstract
Orientadora: Profª. Drª. Leticia Mara Peres Tese (doutorado) - Universidade Federal do Paraná, Setor de Ciências Exatas, Programa de Pós-Graduação em Informática. Defesa : Curitiba, 01/10/2019 Inclui referências: p. 131-136 Área de concentração: Ciência da Computação Resumo: Ontologias sao compostas por restricoes e axiomas que possibilitam descrever formalmente os conceitos, relacionamentos e propriedades de um dominio. A descricao de um conceito pode ter diferentes significados e representacoes atraves da definicao de axiomas. Essa variedade de representacoes permite a ocorrencia de defeitos durante o desenvolvimento de ontologias, que podem ocasionar falhas ou erros inesperados. Outros fatores estao relacionados ao desconhecimento pelo desenvolvedor do uso dos operadores logicos, a enganos em decisoes de modelagem, a inferencias logicas e a definicoes dos requisitos da ontologia. Os defeitos tambem podem ser inseridos pelo desenvolvedor que mesmo tendo conhecimento sobre metodos e tecnicas de modelagem e construcao de ontologias, pode nao conhecer padroes de defeitos que possam ser ocasionados. Com o objetivo de diminuir o numero de defeitos na modelagem de ontologias, e apresentada nesta tese uma contribuicao ao teste de ontologias utilizando o teste de mutacao. Para auxiliar na identificacao dos defeitos que podem ser ocasionados, foi elaborada uma lista contendo os principais defeitos cometidos ao desenvolver ontologias. Estes defeitos foram utilizados como base para o desenvolvimento de um conjunto de 38 operadores de mutacao. Para a aplicacao do teste de mutacao, e apresentado um metodo para a geracao do conjunto de dados de teste, que foi automatizado com a implementacao da ferramenta MutaOnto, assim como a automatizacao da geracao dos mutantes. Para avaliar o conjunto de operadores de mutacao, o metodo de geracao dos dados de teste e a aplicacao do teste de mutacao para ontologias, foram realizados dois experimentos. O primeiro experimento foi realizado com ontologias construidas por desenvolvedores inexperientes, com ontologias que poderiam conter defeitos mais simples e mais faceis de serem identificados. O segundo experimento foi realizado com ontologias selecionadas de uma base online, construidas por especialistas do dominio e do desenvolvimento de ontologias. Dada a experiencia desses especialistas, essas ontologias nao poderiam conter defeitos, ou seus defeitos deveriam ser mais dificeis de serem revelados. Para auxiliar o processo de aplicacao dos dados de teste e analise dos resultados foi utilizada a ferramenta Protégé. O metodo de teste possibilitou revelar 16 tipos de defeitos identificados pelas mutacoes realizadas por 19 operadores de mutacao, sendo que 2 tipos de defeitos foram descobertos apos a aplicacao do teste, pois nao constam no levantamento realizado para a elaboracao dos operadores. Quanto aos operadores de mutacao que nao revelaram algum tipo de defeito, permitiram analisar a qualidade dos dados de teste em identificar as mutacoes e indicar a ausencia do defeito simulado por estes operadores. Palavras-chave: Ontologias, Teste de mutacao, Dados de teste. Abstract: Ontologies are composed of constraints and axioms that make it possible to formally describe the concepts, relationships, and properties of a domain. The description of a concept can have different meanings and representations by defining axioms. This variety of representations enable the occurrence of defects during the development of ontologies that can cause unexpected failures or errors. Other factors are related to the unfamiliarity by the developer of the use of logical operators, mistakes in modelling decisions, logical inferences and definitions of ontology requirements. Defects can also be inserted by the developer, who, although having knowledge about ontology modelling and construction methods and techniques, may not know the patterns of defects that may be caused. In order to reduce the number of defects in ontology modelling, a contribution to the ontology test using the mutation test is presented in this thesis. To help in identifying the defects that may be caused, a list was drawn up containing the main defects committed when developing ontologies. These defects were used as the basis for the development of a set of 38 mutation operators. For the application of the mutation test, a method is presented for the generation of the test data set, which was automated with the implementation of the MutaOnto tool, as well as the automation of the mutant generation. In order to evaluate the set of mutation operators, the method of generating the test data and the application of the mutation test for ontologies, two experiments were performed. The first experiment was carried out with ontologies built by inexperienced developers, with ontologies that could contain simpler and easier to identify defects. The second experiment was carried out with ontologies selected from an online base, built by experts in the domain and ontology development. Given the experience of these experts, these ontologies could not contain defects, or their defects should be more difficult to disclose. The Protege tool was used to assist in the application of test data and results analysis. The test method made it possible to reveal 16 types of defects identified by the mutations performed by 19 mutation operators, and 2 types of defects were discovered after the test, since they are not included in the survey carried out for the elaboration of the operators. As for the mutation operators that did not reveal any type of defect, they allowed to analyse the quality of the test data in identifying the mutations and to indicate the absence of the simulated defect by these operators. Keywords: Ontologies, Mutation test, Test data.
- Published
- 2019
3. Semantic Mutation Test to OWL Ontologies
- Author
-
Porn, Alex Mateus, primary and Peres, Leticia Mara, primary
- Published
- 2017
- Full Text
- View/download PDF
4. Teste de mutação para ontologias OWL
- Author
-
Porn, Alex Mateus, Peres, Leticia Mara, and Universidade Federal do Paraná. Setor de Ciências Exatas. Programa de Pós-Graduação em Informática
- Subjects
Dissertações ,Ciência da computação - Abstract
Orientadora : Profª. Drª. Letícia Mara Peres Dissertação (mestrado) - Universidade Federal do Paraná, Setor de Ciências Exatas, Programa de Pós-Graduação em Informática. Defesa: Curitiba, 15/07/2014 Inclui referências Resumo: Ontologias representam especi_ca_c~oes formais sobre conceitos de um dom__nio e as rela_c~oes entre eles. Elas desempenham um papel fundamental para descrever sem^anticas de dados e atuam como uma espinha dorsal em sistemas baseados em conhecimentos. A diversidade de m_etodos de modelagem de ontologias proporciona que defeitos sejam inseridos durante o desenvolvimento, possibilitando a ocorr^encia de falhas sint_aticas e sem^anticas. M_etodos de testes j_a foram propostos visando revelar falhas em ontologias, entretanto n~ao possibilitam identi_car os defeitos existentes. Com a _nalidade de diminuir o n_umero de defeitos na modelagem de ontologias, _e apresentada neste trabalho a de_ni_c~ao de um processo para a aplica_c~ao do teste de muta_c~ao para ontologias OWL, visando revelar defeitos em ontologias, assim como os seus tipos. Para a gera_c~ao dos mutantes foram propostos 25 operadores de muta_c~ao, por_em, neste trabalho 19 foram aplicados nos modelos em teste. Para a aplica_c~ao do teste foi realizado um estudo de caso no contexto de registros eletr^onicos de sa_ude, onde arqu_etipos openEHR de_nidos na linguagem ADL, foram convertidos para ontologias OWL e submetidos ao processo de teste. Para auxiliar o processo de gera_c~ao dos mutantes e execu_c~ao dos dados de teste, utilizou-se a ferramenta Prot_eg_e de cria_c~ao e edi_c~ao de ontologias. O m_etodo de teste mostrou-se e_caz para determinar a corretude de ontologias, sendo poss__vel gerar um alto n_umero de mutantes e um alto escore de muta_c~ao atrav_es dos operadores propostos e dos dados de teste utilizados, podendo-se identi_car e analisar os defeitos inseridos durante o desenvolvimento. Palavras-Chave: Ontologias. Defeitos de modelagem. Teste de muta_c~ao. Abstract: Ontologies represent formal speci_cations about the domain concepts and the relation between them. They have a fundamental role to describe data semantics and act as a backbone in systems knowledge-based. The insertion of defects during an ontology development is common, causing the occurrence of syntactics and semantics faults. Test methods have been proposed aiming to show faults in ontologies. However these methods does not identify the faults sources. In order to decrease the number of these defects in an ontology design, the de_nition of a process for the application of mutation test to OWL ontologies is presented. It is possible to show existing defects and their types. To generate the mutants, 25 mutation operators have been proposed, but on this paper 19 were applied on models being tested. The application of the test is done by a case study. It was conducted on the context of electronic health records, with openEHR archetypes. These archetypes were written in ADL language, converted to OWL ontologies and submitted to the testing process. It was used Prot_eg_e, a tool of creation and edition of ontologies, to assist the process of mutants generation and test data execution. The test method was e_cacious to determine the correctness of ontologies and was possible to generate a su_cient number of mutants and a high mutation score. Through the proposed operators and the test data used, it was possible identify and analyze the defects inserted during the development. Key-words: Ontologies. Design defects. Mutation test.
- Published
- 2014
5. A SYSTEMATIC LITERATURE REVIEW OF OWL ONTOLOGY EVALUATION.
- Author
-
Porn, Alex Mateus, Gonçalves Huve, Cristiane Aparecida, Peres, Leticia Mara, and Direne, Alexandre Ibrahim
- Subjects
SYSTEMATIC instruction ,ONTOLOGY ,QUALITY control of information storage & retrieval systems ,CONTENT analysis ,COMPUTER performance - Abstract
We present in this paper a systematic literature review, using Kitchenham approach [24], of methods, techniques and tools to evaluate OWL ontologies. We identify methods, techniques, and tools for OWL ontology evaluation, and we classify them based on quality criteria. We conclude that although there is a large number of related studies on ontology evaluation, there are quality criteria which were still unexplored. At the same time, the evaluation process needs computational advances related to tools and processing power. [ABSTRACT FROM AUTHOR]
- Published
- 2016
6. TESTE DE ARQUÉTIPOS
- Author
-
Porn, Alex Mateus, primary and Peres, Leticia Mara, additional
- Published
- 2013
- Full Text
- View/download PDF
7. A Process for the Representation of openEHR ADL Archetypes in OWL Ontologies.
- Author
-
Porn, Alex Mateus, Peres, Leticia Mara, and Del Fabro, Marcos Didonet
- Subjects
NATURAL languages ,DATA analysis ,MEDICAL care ,SEMANTICS ,GENETIC mutation - Abstract
ADL is a formal language to express archetypes, independent of standards or domain. However, its specification is not precise enough in relation to the specialization and semantic of archetypes, presenting difficulties in implementation and a few available tools. Archetypes may be implemented using other languages such as XML or OWL, increasing integration with Semantic Web tools. Exchanging and transforming data can be better implemented with semantics oriented models, for example using OWL which is a language to define and instantiate Web ontologies defined by W3C. OWL permits defining significant, detailed, precise and consistent distinctions among classes, properties and relations by the user, ensuring the consistency of knowledge than using ADL techniques. This paper presents a process of an openEHR ADL archetypes representation in OWL ontologies. This process consists of ADL archetypes conversion in OWL ontologies and validation of OWL resultant ontologies using the mutation test. [ABSTRACT FROM AUTHOR]
- Published
- 2015
- Full Text
- View/download PDF
8. MUTATION TEST TO OWL ONTOLOGIES.
- Author
-
Porn, Alex Mateus and Peres, Leticia Mara
- Subjects
MUTATION testing of computer software ,SEMANTIC Web ,ONTOLOGY ,OPERATOR theory ,COMPARATIVE studies - Abstract
Errors may occur in any human activity. The design of ontologies, as a human activity, may generate defects in the ontology, causing the occurrence of syntactic and/or semantic failures. The mutation test is used in this article to perform the test in OWL ontologies, revealing and reducing the occurrence of failures. A set of new 19 mutation operators for ontologies is proposed to be applied to classes, properties and constraints, using the Protégé tool to assist in the design of ontologies, generation of mutants, execution of test cases and comparison of the results. [ABSTRACT FROM AUTHOR]
- Published
- 2014
9. A Process for the Representation of openEHR ADL Archetypes in OWL Ontologies.
- Author
-
Porn AM, Peres LM, and Didonet Del Fabro M
- Subjects
- Electronic Health Records standards, Humans, Internet, Language, Semantics, Biological Ontologies, Electronic Health Records organization & administration, Natural Language Processing
- Abstract
ADL is a formal language to express archetypes, independent of standards or domain. However, its specification is not precise enough in relation to the specialization and semantic of archetypes, presenting difficulties in implementation and a few available tools. Archetypes may be implemented using other languages such as XML or OWL, increasing integration with Semantic Web tools. Exchanging and transforming data can be better implemented with semantics oriented models, for example using OWL which is a language to define and instantiate Web ontologies defined by W3C. OWL permits defining significant, detailed, precise and consistent distinctions among classes, properties and relations by the user, ensuring the consistency of knowledge than using ADL techniques. This paper presents a process of an openEHR ADL archetypes representation in OWL ontologies. This process consists of ADL archetypes conversion in OWL ontologies and validation of OWL resultant ontologies using the mutation test.
- Published
- 2015
Catalog
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.