The evolution of educational technologies has encouraged the development of adaptive systems that can personalize learning paths. In this context, this paper proposes a method for classifying learning paths on Moodle using syntactic transducers built with the NooJ platform. The proposed method consists of three main steps: (1) Data Extraction from Moodle event logs, (2) Data Classification through NooJ grammars that model activity sequences, and (3) Interactive Dashboard Development to visualize student progress and provide personalized recommendations. To evaluate and validate the method, we applied it to a Moodle course titled “Semantic technology”. The obtained results were highly satisfactory and demonstrate the method’s effectiveness in analyzing learning behaviors and supporting personalized education.