Genetic K-Means Adaption Algorithm for Clustering Stakeholders in System Requirements

A Comparative Systematic Analysis of Stakeholder’s Identification Methods in Requirements Elicitation

“…One possible solution is to cluster many potential stakeholders for the mobile and web-based applications into smaller groups that can be easily managed and analyzed by introducing un-supervising clustering techniques, such as Genetic K-means, Kmeans, etc. [151]. Furthermore, it is mandatory to consider cultural and sociotechnical factors when identifying indirect system stakeholders for marketbased software applications, i-e., face recognition systems, etc., that are currently overlooked, which might have serious consequences [155].…”

“…Next, we use text similarity algorithms, natural language toolkit, and unsupervised learning algorithms, such as KMean, genetic KMean, etc. to identify similar stakeholders in the social media platforms [151]. Finally, we can identify similar stakeholders if slightly different end-users names have similar email login names.…”

Salience Based Stakeholders Selection Keeping Coverage in the Next Release Problem Solution

“…This identification has a clear impact on the problem of requirement selection by reducing the workload; as with a smaller group of stakeholders, we can identify a set of requirements that satisfies the demands of the overall community of stakeholders. Other works, such as [20,21,32,9,36] focus on the impact that stakeholder weight has on requirement selection, but do not accomplish any stakeholder classification task. Instead, they embedded all stakeholder data into the requirement selection problem.…”

“…The first refers to our estimates about power, legitimacy, and urgency, which have been derived from the available data records. Being this adaptation of the hand a possible threat, it has been based, respectively, on the ground truth for power, on a recommendation network for legitimacy (which has been validated in several works [39,21,35,38]), and simple votes count for urgency. This mapping is adequate within the scope of our work and captures the meaning that stakeholder theory defines for each salience component [24].…”

“…However, they left aside the stakeholder salience theory, which characterises stakeholders based on power, legitimacy, and urgency. Examples of this are proposals to identify and prioritise stakeholders in large-scale software projects with the aim of prioritising requirements [20,21], where stakeholder weights are estimated using only social network recommendations, that is, stakeholder legitimacy; or those that use the complete stakeholder preference matrix for requirements to apply genetic clustering to find convergence between the preferences of stakeholders to define groups of stakeholders [32].…”