Software Risk Modeling by Clustering Project Metrics
Ching-Pao Chang · International Journal of Software Engineering and Knowledge Engineering · 2015
Together with the development and integration of software technologies, an increase in the complexity of the software development environment has made identifying software risks challenging. Identifying software risks, which is a critical activity in project management, is challenging because numerous factors may affect software projects. In this paper, an approach to identify software-risk items is proposed in which data collected from past software projects are mined to construct software-risk models. The prediction models obtained can be used to identify potential software risks for subsequent software projects. The advantage of the proposed approach is that the software-risk models can be constructed at an early stage in software projects to facilitate the planning of methods to mitigate software risks. The proposed approach is applied to a business project to demonstrate how software risk items can be identified.