An academic project; applying MDE principles in order to mine object oriented projects.
With the rise in popularity of object-oriented programs, it is more vital than ever to learn about data mining techniques to discover knowledge, patterns, and relationships between elements that can be used in the future to produce accurate predictions. In fact, achieving a good performance of a given model is determined not only by the model and hyperparameters, but also by how different types of variables are processed, encoded and fed into the model. We propose to implement an approach based on Model driven engineering principles to mine code repositories and extract knowledge in order to provide ready to use data for graph based machine learning algorithms. To this end, we are planning to use encoding algorithms to learn domain concepts and their relationships from identifiers used in diagrams reverse engineered from the code.
Keywords:
MDE, Reverse Engineering, Transformation, ATL, EGL
process:
For the rest of the report, feel free to contact me: meriem.ben.chaaben@umontreal.ca