In this paper a novel approach to facilitate the reuse of units of learning (UoLs) is presented. Typically, e-learning platforms do not provide the means to retrieve designed UoLs in a standardized format to be reused in a different platform, but they have in common that the students and teachers interaction with the system is logged to files. Taking this into account, we propose to use these logs and apply a three steps re-engineering approach to translate these UoLs into an accepted educational modelling language, specifically IMS LD. In the first step, the sequence of activities and their functional dependencies are learned by a process mining algorithm. In the second step, another algorithm analyses the variables and their value change in order to learn the adaptation rules that may have been defined in the UoL. And finally, in the last step the inferred process structure and rules are matched with the typical structure of activities, acts, and plays defined by IMS LD.
Keywords: IMS LD, process mining, adaptive rules mining, learning analytics, learning flows reengineering