Learning from evolution history to predict future requirement changes

摘要: Managing the costs and risks of evolution is a challenging problem in RE community. The challenge lies difficulty analyzing assessing proneness to requirement changes across multiple versions, especially when scale requirements large. In this paper, we define series metrics characterize historic information, propose novel method for predicting that are likely evolve future based on metrics. We apply prediction analyze product updates history through case study. empirical results show can provide tradeoff solution narrows down scope change analysis small set requirements, but it still retrieve nearly half changes. indicate defined sensitive evolution, reach valuable outcome engineers balance their workload risks.