Uso de predição de defeitos aplicada ao processo de identificação e recomendação para priorização de bugs

• Autor(a) principal: Pontes, Alan Bonifácio de
• Data de Publicação: 2017
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFPB
• Texto Completo: https://repositorio.ufpb.br/jspui/handle/123456789/12907
• topic: Predição de defeitos de software; Testes de software; Inteligência artificial; Engenharia de software; Software defect prediction; Software testing; Artificial intelligence; Software engineering; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

The use of computational techniques has been increasing considerably, thus providing a progressive development of science and scientific research in all areas of knowledge. Computer systems significantly assist in this process and the quality of these systems becomes an essential factor, since software are more and more complex. Also, in order to achieve the desired quality, the stage of software testing becomes required and it is essential that it be executed in the most optimized way possible, for it is a consensus that this step requires both time and money. With this, defect prediction techniques have become a great ally in discovering software defects. In this dissertation we relied on the many defect prediction techniques available as well as on computational metrics, and from the confusion matrix we were able to indicate the most promising models to predict software defects and also to aid in indicating possible bug prioritization recommendations. In this work, through various simulations, it is verified the validation of the models that are created from the older versions of the UFPB’s academic system, applying such models in the most recent versions. In other words, earlier versions of the system should indicate possible errors in future versions of the same system. In addition, automated indication of bug priority is also possible. What is intended is to give support to the team so it can prioritize the preparation and execution of test scripts in the next versions, as well as indicate, based on historical data, the priority of the error found.