III: Small:Using Data Mining and Recommender Systems to Facilitate Large-Scale Requirements Processes

与需求定义相关的问题导致了大量的项目失败，并转化为大量的资金浪费。在许多情况下，这些问题的根源是，在征求利益攸关方的需求这一人力密集型任务方面存在不足，以及随后在将这些需求转化为一套明确阐述和排定优先次序的要求方面存在问题。这些问题在联邦调查局虚拟案件档案或美国宇航局空间站等大型项目中尤为明显，这些项目的知识分散在数千个不同的利益相关者中。一方面，在启发和确定优先级的过程中，尽可能多的人参与是可取的，但另一方面，这可能很快导致信息和意见的混乱过载。根据这项拨款提出的工作将开发一个新的框架，利用数据挖掘和推荐系统技术来处理和分析大量的非结构化数据，以促进大规模和广泛包容的需求过程。该建议是基于这样的观察，即许多大型工业和政府项目的需求获取过程本质上是数据驱动的，因此可以受益于基于数据挖掘和用户建模技术的计算机支持的工具。智力优势-拟议中的研究将导致一个强大的需求启发框架和相关的工具库，可用于增强功能的维基，论坛和专门的管理工具中使用的需求域。具体来说，这项研究将提高需求聚类技术，结合先验知识和用户派生的约束。将设计一个情境化推荐系统，以便于将利益相关者适当地安排到聚类阶段生成的需求讨论论坛中。更广泛的影响拟议的工作有可能对开发知识产权密集型系统的组织产生广泛的影响。由于与西门子和谷歌等组织的合作计划是本研究的一个组成部分，因此可以预期技术转让。将专门为需求工程和推荐系统课程编制教材，并将广泛分发。关键词：推荐系统;数据挖掘;聚类;需求工程;需求获取。

Problems related to requirements definitions account for numerous project failures and translate into significant amounts of wasted funds. In many cases, these problems originate from inadequacies in the human-intensive task of eliciting stakeholders' needs, and the subsequent problems of transforming them into a set of clearly articulated and prioritized requirements. These problems are particularly evident in very large projects such as the FBI Virtual Case File or NASA's Space Station, in which knowledge is dispersed across thousands of different stakeholders. On one hand, it is desirable to include as many people as possible in the elicitation and prioritization process, but on the other hand this can quickly lead to a rather chaotic overload of information and opinions. The work proposed under this grant will develop a new framework that utilizes data mining and recommender systems techniques to process and analyze high volumes of unstructured data in order to facilitate large-scale and broadly inclusive requirements processes. The proposal is based on the observation that the requirements elicitation process of many large-scaled industrial and governmental projects is inherently data-driven, and could therefore benefit from computer-supported tools based on data mining and user modeling techniques. INTELLECTUAL MERIT The proposed research will lead to a robust requirements elicitation framework and an associated library of tools which can be used to augment the functionality of wikis, forums, and specialized management tools used in the requirements domain. Specifically, this research will enhance requirements clustering techniques by incorporating prior knowledge and user-derived constraints. A contextualized recommender system will be designed to facilitate appropriate placement of stakeholders into requirements discussion forums generated in the clustering phase. BROADER IMPACT The proposed work has potential for broad impact across organizations that develop stakeholder-intensive systems. Technology transfer can be expected due to collaborations with organizations such as Siemens and Google planned as an integral part of this research. Educational materials will be developed specifically for requirements engineering and recommender systems courses, and will be broadly disseminated. Key Words: Recommender systems; Data mining; Clustering; Requirements engineering; Requirements elicitation.