Arabic Non-Functional Requirements Extraction Using Machine Learning

Non-Functional Requirements (NFR) are a set of quality attributes that software must have, such as security, reliability, and performance. Extracting NFR from software requirement specifications can help developers deliver quality software that meets users' expectations. However, since functional and non-functional requirements are mixed in the same SRS, it requires a lot of human effort to distinguish them. While many studies have proposed English language requirements extracting techniques, there is a lack of research in Arabic requirements extracting, as well as a lack of publicly available Arabic datasets in this field. In this study, we propose an automatic NFR extraction method for quality software development by combining machine learning and feature extraction techniques. Also, we will collect an Arabic dataset for requirements. This study aims to help software engineers save time, reduce costs and effort in the manual extraction process, and make the requirements engineering phase more efficient. Additionally, it provides new research areas in this field.