Enhancing Learning Analytics in Open-Source Software Mailing Archives using Machine Learning and Process Discovery Techniques.

Existing evidence indicates that Free/Libre Open-Source Software (FLOSS) ecosystems offer extensive learning opportunities. Community members actively participate in various activities, both during their interactions with peers and while utilizing these environments. Given that FLOSS repositories contain valuable data on participant interactions and activities, our study focuses on analyzing knowledge exchange and interactions within emails to track learning activities across different phases of the learning process, with a focus on the first phase (Initiation). In this paper, we leverage Natural Language Processing (NLP) and Machine Learning (ML) techniques within a process mining framework. Specifically, we employ NLP techniques to analyze the contents of emails and messages exchanged in these FLOSS repositories to generate event logs for the purpose of modeling learning patterns. Subsequently, we construct corresponding event logs, which serve as input to Disco, the process mining tool, for learning process discovery in these environments. The output comprises visual workflow nets that we interpret as representations of learning activity traces within FLOSS, capturing their sequential occurrences. To enhance the understanding of these models, we incorporate additional statistical details for contextualization and description. This approach enables a nuanced exploration of learning dynamics within FLOSS environments, emphasizing the role of NLP and ML in uncovering valuable insights on how FLOSS participants acquire and exchange knowledge. [ABSTRACT FROM AUTHOR]