Classificando a qualidade de requisitos de sistemas com a utilização de algoritmo de redes bayesianas

During the life cycle of a software project it is common that problems with the system requirements are identified, causing problems such as: increase in project cost, delay in deliver, rework in codes, among others. In this scenario companies are looking for alternatives to improve this process, such as the adoption of new software development paradigms, but they can not completely eliminate the uncertainty regarding the quality of the requirement. This work seeks to demonstrate how the use of bayesian networks, whose pilot prototype had an accuracy of 72.13%, can facilitate the identification of the quality of the requirements in the initial phase of the project and leads to minimize the impacts on the project budget