Towards aspect based requirements mining for trace retrieval of component-based software management process in globally distributed environment

Stakeholder satisfaction is a significant aspect of component-based product development. Satisfaction level of stakeholder varies due to diverse reviews and perspective about components functionalities. The reviews and perspective create ambiguities and misunderstanding during management of components requirement from specification to linking requirements that lead to product failures. The improper components management increases efforts and errors when component’s stakeholders and development team is located in a globally distributed environment. The major issues of distributed component-based systems, are control, communication, coordination, and semantical analysis of different reviews and perspectives. As requirements of components is elicited and developed at different locations which created ambiguities and irrelevancy during components integration. Therefore, in this study, we proposed a framework to improve the management process of components requirement in a distributed environment. To reduce ambiguities and incompleteness among requirements, aspect based sentiment analysis has been utilized for each stakeholders’ reviews and perceptive individually. On the other hand, to reduce involvement of stakeholder and efforts in components prioritization and linking processes, we adapted cased based reasoning method and decision tree-based classification of requirements, respectively. The performance of the proposed framework has been evaluated through an experimental approach in order to compare it with current practices i.e. Random selection and expert based. The findings described that the accuracy of component management in global development increases with proposed framework. Further, results show that there is an increase in product quality with decrease in irrelevancy and redundancy in stakeholders’ aspects and priority.