This repository is the reproduction package for the experiments from the paper "Schindler, C and Rausch, A. Formal Software Architecture Rule Learning: A Comparative Investigation between Large Language Models and Inductive Techniques. 2024."
We provide the knowledge base (first-order logic facts) and knowledge graphs, extracted from the source code of Teammates (https://github.com/sebastianherold/SAEroConRepo/tree/master/systems/TEAMMATES).
For the conducted experiments with Inductive Rule Learning (IRL) tools and Large Language Models (LLMs) we provide the experiment results in form of logs of the tool runs and chat interactions with the LLMs. For each IRL tool, we also include the tool configurations and instructions on how the tools were used.