Salience-based stakeholder selection to maintain stakeholder coverage in solving the next release problem

Abstract: Stakeholders quantification plays a basic role in selecting the appropriate requirements because their judgement is a major criteria since not all of them have the same importance. Original proposals quantified stakeholders assigning them a weight. Nonetheless, actual projects manage a numerous stakeholders community hindering the inclusion of all their weights. This work proposes grouping strategies as means to reduce the number of stakeholders to manage in requirements selection keeping a proper coverage (i.e. howthe selection fulfils stakeholder demands). Our approach is based on stakeholders' salience, defined in terms of power, legitimacy and urgency attributes. Diverse strategies are applied selecting important stakeholders groups in a specific project. We use k-means}, k-medoids and hierarchical clustering, after deciding the number of cluster (4 and 3) based on validation indices. Either for all the stakeholders and each important group several requirements selection optimization problems have been solved. Tests find no significant differences for coverage when important stakeholders are filtered using clustering, regardless of the technique and number of groups, with a reduction between 66.32% to 87.75% in the number of stakeholders being considered. Applying clustering methods on data obtained from a real-world project is useful to identify the group of important stakeholders. The number of groups suggested matches the stakeholders theory and the coverage values in the requirements selection is kept.