Uma revisão sistemática da literatura sobre comitês de classificadores para predição da estimativa do esforço de equipe de software

Abstract

The Prediction of software team effort estimation is one of the key aspects of a project's life cycle. Predicting this estimate remains a challenging step, and various techniques have been used over the years, with the use of classifier committees standing out as a promising technique. Therefore, the aim of this study was to conduct a systematic literature review (SLR) on the use of classifier committees for predicting software team effort estimation. The Kitchenham (2009) study methodology was used to conduct the SLR process, and the search sources used to gather the studies were IEEE Xplore, Scopus, ACM, and Web of Science, with the studies published from 2019 to 2023. In the collected studies, the Desharnais dataset was predominantly used, and the Weka tool/framework was the most commonly employed.The most applied evaluation metric for software effort estimation prediction was PRED, and the most addressed validation technique was Cross Validation. In the assessment of study quality, the study by Rai, Kumar, and Verma (2021) obtained the highest total score. Each machine learning technique used in the studies has its strengths and weaknesses, so it was not possible to determine the best technique.