Abstract: The semantic description requirements of large-scale users were ambiguous and uncertain,and the correlation between the requirements was complex.In this situation,a modeling and analysis method for semantic description requirements of large-scale users was proposed.On the basis of analyzing the forms and types of semantic description requirements,a requirement model was constructed and a two-stage requirement modeling process was formulated.The requirements were synthesized first.The non-parametric requirements were clustered based on the semantic similarity of perceptual words,and the parametric requirements were clustered based on the similarity of trapezoidal fuzzy numbers corresponding to parametric viewpoints.The semantic description requirements were transformed into standard requirements.The correlations between requirements were evaluated by a two-stage assessment method,the quantities of requirements were corrected and the importance of requirements was calculated accordingly.Key requirements were identified according to the importance of requirements,and users requirement scheme was output.A case study of the modeling and analysis process of large-scale user semantic description requirements for a hand-held vacuum cleaner was conducted.The case study results showed that the method was feasible.

Key words: large-scale users, semantic description requirements, requirement modeling, semantic similarity, trapezoidal fuzzy numbers

摘要： 针对大规模用户提交的语义描述需求存在模糊性、不确定性、需求关系复杂等问题,提出一种大规模用户语义描述需求建模与分析方法。在分析语义描述需求形式和类型的基础上,构建需求模型并制定两阶段需求建模过程。首先合成需求,基于感性词语义相似度聚类非参数型需求,基于参数观点对应的梯形模糊数相似度聚类参数型需求,将语义描述需求转化为标准的需求项;然后,采用两轮评估的方法评估需求间的相关关系,并据此修正需求量和计算需求重要度,根据需求重要度识别关键需求,输出用户群体需求方案。以手持吸尘器的大规模用户语义描述需求建模分析过程为例对该方法进行应用研究,证明了方法的可行性。

关键词: 大规模用户, 语义描述需求, 需求建模, 语义相似度, 梯形模糊数