Learning from evolution history to predict future requirement changes

ISCAS OpenIR

	Learning from evolution history to predict future requirement changes
	Shi, Lin (1); Wang, Qing (1); Li, Mingshu (1)
	2013
会议名称	2013 21st IEEE International Requirements Engineering Conference, RE 2013
页码	135-144
会议日期	July 15, 2013 - July 19, 2013
会议地点	Rio de Janeiro, Brazil
收录类别	EI
出版地	IEEE Computer Society, 2001 L Street N.W., Suite 700, Washington, DC 20036-4928, United States
ISBN	9781467357654
部门归属	(1) Laboratory for Internet Software Technologies, Institute of Software, Chinese Academy of Sciences, Beijing, China; (2) University of Chinese Academy of Sciences, Beijing, China; (3) State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences, Beijing, China
摘要	Managing the costs and risks of evolution is a challenging problem in the RE community. The challenge lies in the difficulty of analyzing and assessing the proneness to requirement changes across multiple versions, especially when the scale of requirements is large. In this paper, we define a series of metrics to characterize historic evolution information, and propose a novel method for predicting requirements that are likely to evolve in the future based on the metrics. We apply the prediction method to analyze the product updates history through a case study. The empirical results show that this method can provide a tradeoff solution that narrows down the scope of change analysis to a small set of requirements, but it still can retrieve nearly half of the future changes. The results indicate that the defined metrics are sensitive to the history of requirements evolution, and the prediction method can reach a valuable outcome for requirement engineers to balance their workload and risks. © 2013 IEEE.; Managing the costs and risks of evolution is a challenging problem in the RE community. The challenge lies in the difficulty of analyzing and assessing the proneness to requirement changes across multiple versions, especially when the scale of requirements is large. In this paper, we define a series of metrics to characterize historic evolution information, and propose a novel method for predicting requirements that are likely to evolve in the future based on the metrics. We apply the prediction method to analyze the product updates history through a case study. The empirical results show that this method can provide a tradeoff solution that narrows down the scope of change analysis to a small set of requirements, but it still can retrieve nearly half of the future changes. The results indicate that the defined metrics are sensitive to the history of requirements evolution, and the prediction method can reach a valuable outcome for requirement engineers to balance their workload and risks. © 2013 IEEE.
语种	英语
内容类型	会议论文
URI标识	http://ir.iscas.ac.cn/handle/311060/16689
专题	中国科学院软件研究所
推荐引用方式 GB/T 7714	Shi, Lin ,Wang, Qing ,Li, Mingshu . Learning from evolution history to predict future requirement changes[C]. IEEE Computer Society, 2001 L Street N.W., Suite 700, Washington, DC 20036-4928, United States,2013:135-144.