Automated Features and Requirements Identification for Improving CPS Software Reuse using Topic Modeling

Software reuse is a common practice in software development due to its ability to reduce development costs, accelerate time to market, and mitigate the risks associated with building a new system from scratch. Cyber-physical systems (CPS) are no exception to this trend, with many existing reusable CPS software available in public repositories such as GitHub. However, identifying CPS software features and requirements from this legacy code is challenging, as it requires developers' domain knowledge to understand the system's functionality and configurations. Moreover, in many legacy software, the original design and documentation may be incomplete or unavailable, making it more challenging to extract these features and requirements manually. To address this challenge, we propose an automated approach for identifying CPS software features and requirements using topic modeling and code analysis techniques. We evaluate our approach compared to manual and rule-based approaches, and the results show that it outperforms these approaches in terms of precision, recall, and F1 score.