10.29327/1407529.27-24
Despite the increasing development of Machine Learning (ML) applications, Requirements Engineering (RE) activities face challenges in this new data-intensive paradigm, e.g., the high dependence on data availability and quality and the continuous adaptation to changing environments. In this context, we have identified a lack of an integrated view of the RE process of ML applications in the literature. This paper proposes an RE process tailored to ML application projects, covering requirements elicitation, analysis, specification, validation, and management. The process development methodology includes the ISO/IEC 12207 standard and Design Thinking. Our solution combines problem study and formulation supported by the state of the art in research. The solution proposal follows three evaluation formats: laboratory, static, and dynamic validation. We expect the process to improve RE practice and thus improve ML-based systems development with higher quality deliveries and easier maintenance.
Requirements Engineering; Machine Learning; ISO/IEC 12207; Software Process; Design Thinking; Technology Transfer Model
@inproceedings{wer202430,
author = {Martins, M. C. and Kudo, T. N. and Bulcão-Neto, R. F.},
title = {A Requirements Engineering Process for Machine Learning Innovation Projects},
booktitle = {Proceedings of the WER2024-27th Workshop on Requirements Engineering, Buenos Aires, Argentina},
year = {2024},
issn = {2675-0066},
isbn = {978-65-01-06131-3},
doi = {10.29327/1407529.27-24}
}