Critérios para priorização de estudos primários identificados por snowballing com conjunto inicial gerado por string de busca
Ano de defesa: | 2017 |
---|---|
Autor(a) principal: | |
Orientador(a): | |
Banca de defesa: | |
Tipo de documento: | Dissertação |
Tipo de acesso: | Acesso aberto |
Idioma: | por |
Instituição de defesa: |
Universidade Federal de São Carlos
Câmpus São Carlos |
Programa de Pós-Graduação: |
Programa de Pós-Graduação em Ciência da Computação - PPGCC
|
Departamento: |
Não Informado pela instituição
|
País: |
Não Informado pela instituição
|
Palavras-chave em Português: | |
Palavras-chave em Inglês: | |
Área do conhecimento CNPq: | |
Link de acesso: | https://repositorio.ufscar.br/handle/20.500.14289/13538 |
Resumo: | Background: In any area of science, all proposed technology should be evaluated and characterized so that the community can understand when, where and how to apply it. Particularly in the Software Engineering field, the concern with this characterization and the characterization of the technology is known as Evidence-Based Software Engineering (EBSE). Secondary studies contribute to this characterization because they use more rigorous research methods. Snowballing is a technique to identify relevant primary studies in secondary studies. It is well accepted by the community but may require a great deal of effort in its implementation depending on the guidelines adopted and on the number of existing relevant primary studies. Objective: To propose criteria for prioritizing primary studies for applying snowballing using the results obtained from search strings as seed aiming at improving the efficiency when applying the technique, providing computational support for its application. Method: Four criteria were proposed for prioritization of primary studies and algorithms were developed that allow computational support for applying the criteria. The evaluation was carried out by means of experimental studies, being an exploratory study and two experiments. Results: The best results were obtained by combining the four proposed criteria, resulting in a precision of 96.46% and a recall of 51.24%, with an average error rate of 20%. Algorithms created for reference extraction and similarity calculation presented satisfactory results. Conclusion: Based on the obtained results, there are indications that the proposed prioritization criteria, which make use of the reference extraction and similarity algorithms developed and implemented in the StArt tool, contribute to increase the efficiency of the snowballing technique without having a significant loss of quality of the results. |