A Data-Driven Framework for Automated Requirements Elicitation from Heterogeneous Digital Sources
Increased digitalization and the pervasiveness of Big Data, along with vastly improved data processing capabilities, have led to the consideration of digital data as additional sources of system requirements, complementing conventional stakeholder-driven
- PDF / 1,069,428 Bytes
- 15 Pages / 439.37 x 666.142 pts Page_size
- 50 Downloads / 173 Views
Abstract. Increased digitalization and the pervasiveness of Big Data, along with vastly improved data processing capabilities, have led to the consideration of digital data as additional sources of system requirements, complementing conventional stakeholder-driven approaches. The volume, velocity and variety of these digital sources present numerous challenges which existing system development methods are unable to manage in a systematic and efficient manner. We propose a holistic and data-driven framework for continuous and automated acquisition, analysis and aggregation of heterogeneous digital sources for the purposes of requirements elicitation and management. The proposed framework includes a conceptualization in the form of a meta-model and a high-level process for its use; the framework is illustrated in a real case of an enterprise software. Keywords: Enterprise modeling · Data-driven requirements engineering · Meta-model · Natural language processing · Machine learning · User stories
1 Introduction Requirements elicitation is traditionally stakeholder-driven, where the primary sources of information are generated through interviews with business experts [1, 2]. Due to business and technology changes, intensive industry competition, as well as changing needs and expectations of customers, it is not possible to fully determine the desired behavior of a software system prior to development; instead, it is often preferable to build an initial solution fast, which is later evolved over time. Although agile methodologies have contributed to this by introducing development practices that facilitate interactive and timely software delivery [3], it has become relevant to enable the means for eliciting requirements from a broader spectrum of sources [4]. Increased digitalization has led to continuous generation of large amounts of digital data, both in organizations and in society at large. In the requirements engineering (RE) community, there has been a growing interest in considering digital data as additional sources for requirements acquisition [4, 5]. Considering digital data for requirements acquisition is important for a number of important reasons: (i) it allows for increased © IFIP International Federation for Information Processing 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved J. Grabis and D. Bork (Eds.): PoEM 2020, LNBIP 400, pp. 351–365, 2020. https://doi.org/10.1007/978-3-030-63479-7_24
352
A. Henriksson and J. Zdravkovic
scalability and to consider a wider scope of stakeholders, including potentially large and dispersed user bases (ii) it facilitates increased automation, both of elicitation and subsequent RE activities (iii) it makes continuous elicitation possible, supporting software evolution through more frequent software improvements. There is a wide range of digital sources that may be exploited, but recently increasing attention has been given to sources that are more dynamic – i.e. continuously generating large amounts of data – and often non-intended, i.e. d
Data Loading...
