An Approach for Software Development Effort Estimation Using ChatGPT

Effort estimation poses a significant challenge in software development as it encompasses the determination of the necessary time and resources for the completion of a project. Several approaches have been proposed to estimate the effort in software development, but they often exhibit various limitations which can be grouped into three main categories. Firstly, these methods heavily rely on expert judgment, introducing subjectivity and variations in estimation. Secondly, they can be complex to comprehend and implement, requiring a deep understanding of software metrics and project attributes. Finally, they usually need manual effort in collecting and analyzing data, leading to potentially time-consuming tasks and the possibility of errors. This paper presents a solution that addresses these challenges by utilizing established conceptual models like Business Process Management Notation (BPMN) to develop a comprehensive system description through the extraction of entities and their relationships as semantic triples. This approach lays the foundation for identifying conceptual micro-services, enabling a precise breakdown of system functionalities, and incorporates prompt engineering with ChatGPT for suitable effort estimation of each micro-service.