【摘要】 为了实现大数量软件需求的优选,提出了一种基于自然语言理解的需求聚类和需求优选方法。该方法首先基于自然语言处理技术对需求进行深层次语言理解,找到相同语义的需求并只保留一条。随后,采用基于超图分割的需求聚类方法对大量需求进行聚类。最后,提出了需求优选目标函数,该目标函数根据需求聚类结果为需求优先级打分,并在综合考虑需求与类别的相似性、类别的权重以及聚类标准的权重的基础上,最终给出一个合理的需求优选结果。在大规模真实的需求集合上进行了实验,实验结果表明:基于自然语言处理技术的需求聚类性能优异;需求聚类对后续的需求优选起有着非常重要的作用;基于超图分割的需求聚类,以及在此基础上提出的需求优选函数相对于基线方法有着明显的提高。更多还原

【Abstract】 To realize the optimal selection of great quantities of software requirements,a new requirement prioritization approach based on natural language understanding is presented.The approach is described below.Firstly,it uses the natural language processing technique to deeply understand requirements from the angle of language and find the requirements of same semanteme,with one of them being reserved.Secondly,it adopts a requirement clustering method based on hypergraph partitioning to cluster quantities of requirements.Finally,it presents an object function for optimal selection of requirements.The function gives marks for requirement priorities according to requirement clustering results,and at last,gives a rational requirement clustering results,and at last,gives a rational requirement prioritization result based on the comprehensive consideration of the similarty between requirement and classification,the weight of classification,and the weight of clustering standard.The results of the experiment performed on a real large-scall requirement set show that the performance of the requirement clustering based on the natural language processing is excellent,the requirement clustering is very important to the follow-up requirement prioritization,and the requirement clustering based on hypergraph partitioning and the function for requirement prioritization outperform the base line method.更多还原