CITI has stopped operations in 2014, to co-launch NOVA LINCS THIS SITE IS NOT BEING UPDATED SINCE 2013
citi banner
Home Page FCT/UNL UNL
  Home  \  Graduation Activities  \  Post-Graduation Page Login  
   
banner bottom
File Top
Mineração de Repositórios de Software
MSc Post-Graduation
Abstract:

The discovery, or confirmation, of trends and patterns in the evolution of software systems has been giving importance to mining software repositories. Software Engineering uses specific approaches to mining data from software construction, such as: source code, version history (logs), bug reports (defect tracking), among others. Members of the development team are essential during the preparation and maintenance of software. New integrated software tools related to development activities have been adopted to optimize their work and these have allowed data (such as evolution requests, version control repositories, etc.) to be stored automatically. These data can be recovered and analyzed in order to obtain important information for facilitating software process improvement. Mining the repositories of the development process allows detecting trends and patterns both in the development process and in the developed artifacts, making it an important tool to support the management of the software development process. In this study we use information contained in repositories systems of evolution requests so as to create predictive models of the distribution of those requests throughout the time. These types of models are useful to facilitate the management of the development and maintenance of software, since they will be able to predict periods with a great deal of orders, in contrast to other periods in which there will be fewer, and that information is relevant to the allocation of human resources to the process of development and maintenance. The approach to be used aims to study the most appropriate types of models, according to existing historical data and the pattern of versions which the repository deals with. In particular, we want to know:  Is the choice of the "best" model relatively stable, or too volatile? The implication is that we may have to update models very often or not at all.  Do the models incorporating seasonal information become dominant? If so, how much historical data is required for the seasonal information to be relevant?


Start Date: 2012-11-08

End Date: 2013-11-27


URL: http://ctp.di.fct.unl.pt/~mgoul/papers/2013/GoretiSilva2013MSc.pdf




Post-Graduation Student / Researcher / Professor:
  • Goreti Silva ( Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa )

Post-Graduation Supervisor(s):

Post-Graduation Jury:
File Bottom