In this paper, we formulated, designed, implemented and evaluated a model used for classifying stakeholders' requirements that are specified for web application development. The study employed both qualitative and quantitative research approaches in a case study. Requirements were elicited from stakeholders using the interview approach. This involved speaking with the stakeholders directly via groupware and asking them questions about their specific needs that are relevant to the development of web application. In particular, 10 customers of Procrea8 Technology Solution Limited and 9 developers were used as respondents. An interactive genetic algorithm was used to formulate the model. The design was specified using the Unified Modeling Language (UML) tool, and implemented using specified web technology tools. The model was evaluated for completeness and consistency using recall and precision as parameters. The results showed that a list of ordered requirements was produced based on the stakeholders' priorities inputted into the model. The output indicated the order of priorities finally assigned to each of the requirements. The evaluation revealed that the model is effective, efficient, userfriendly, reliable (with 96.3% accuracy), scalable (prioritized over 500 requirements), less timeconsuming (prioritizing over 500 requirements) and able to update ranks whenever changes occur automatically. Also, the model evaluation indicates 97.1% precision (consistency), and 96.0% recall (completeness). The study shows that requirements engineers could use the model to collate stakeholders’ requirements from wide geographical locations.
Keywords: Requirements analysis, requirement prioritization, requirements engineering, web application, requirement specification.
A COSMIC-FFP APPROACH TO ESTIMATE WEB APPLICATION DEVELOPMENT EFFORT
