A semi-automated requirements prioritisation technique for scalable requirements with stakeholder quantification and prioritisation

One of the gatekeepers of quality software systems is requirements prioritisation (RP) that is often used to select the most important requirements as perceived by system stakeholders. RP is considered as a vital role in ensuring the development of a quality system with defined constraint. Stakeholder quantification and prioritisation (SQP) is executed to quantify and prioritise the stakeholders of the system based on their impacts. The SQP plays a crucial role in identifying and selecting the most essential requirements to produce a successful system. Thus, this research mainly focuses on the RP and SQP domains. Although, the useful of the existing RP and SQP techniques, a close look discloses that these techniques face key challenges with respect to the scalability, shortage of SQP process, lack of low SQP implementation detail with respect to the non-existence of attributes measurement criteria and heavily need of highly professional human intervention in quantifying and prioritising the participating stakeholders and specifying priority value of each requirement in RP process, and lack of automation along time consumption in performing the SQP and RP processes. Hence, a new semi-automated scalable prioritisation technique (SRPTackle) integration with a new SQP technique (StakeQP) are proposed to address the reported key limitations. The StakeQP introduces new low-level implementation details to perform SQP automatically. The StakeQP is on the basis of the newly proposed new measurement criteria for each SQP attribute and using the multi-attribute decision-making method, namely, technique of order preference similarity to the ideal solution (TOPSIS). Furthermore, the proposed SRPTackle is based on the combination of the proposed StakeQP technique, the constructed requirement priority value formulation function and the employing of classifying algorithm (K-means and K-means++) and binary search tree. The effectiveness of SRPTackle and StakeQP are evaluated using a benchmark dataset of the actual software project (RALIC). Experimental implementation of the proposed StakeQP technique and comparative analysis against the existing SQP techniques have been conducted in order to evaluate the StakeQP performance. On other hand, seven experiments are conducted using the large sets of requirements with purpose of assessing the SRPTackle and comparing the SRPTackle performance results with other alternative techniques. The experiments show that StakeQP can produce accurate result of 89.69 %, while accuracy results of the SRPTackle are 93.0% and 94.65% as minimum and maximum accuracy, respectively, which are better than other existing SQP and RP techniques. Also, the findings demonstrate that the StakeQP and SRPTackle perform the SQP and RP process, respectively with less time consumption and are more effective in addressing the reported key limitations compared with other alternative techniques. Future research can dig deeper in improving the SRPTackle and StakeQP performance in terms of catering the requirements independencies and stakeholder classifications, respectively, along with extending the implication of the StakeQP and SRPTackle with different dataset of global software projects practices for better applicability