Nhatos: um modelo para recomendação de requisitos baseado em históricos de contextos

Abstract

The software project management environment has numerous uncertainties. During the life cycle of projects, new requirements, and requests for changes occur at all times, making their management strategic. Failures that occur during requirements management impair the chances of successful projects. However, proper requirements management reduces the risk of deviations from planning about the time, cost, and quality of the project. Field research was conducted involving project professionals to collect the teams’ perception of the application of a proactive computational model to support software teams during their decision making. In this sense, this qualification presents a model called Nhatos. Nhatos contributes to reducing the probability of failure of projects through the recommendation of requirements. Nhatos follows ubiquitous computing concepts, since it recommends requirements through the analysis of historical project contexts, considering the advances in their life cycles. The model was exposed to an industry environment and evaluated using a historical basis containing 183 projects. After analyzing different scenarios, Nhatos achieved a correct percentage of 83.04% of the recommendations made.