An effective approach for software project effort and duration estimation with machine learning algorithms

摘要: Abstract During the last two decades, there has been substantial research performed in field of software estimation using machine learning algorithms that aimed to tackle deficiencies traditional and parametric techniques, increase project success rates align with modern development management approaches. Nevertheless, mostly due inconclusive results vague model building approaches, are few or none deployments practice. The purpose this article is narrow gap between up-to-date implementations within organisations by proposing effective practical deployment maintenance approaches utilization findings industry best practices. This was achieved applying ISBSG dataset, smart data preparation, an ensemble averaging three (Support Vector Machines, Neural Networks Generalized Linear Models) cross validation. obtained models for effort duration intended provide a decision support tool develop implement systems.