The code of our article: "DF4RT: Deep Forest for Requirements Traceability Recovery between Use Cases and Source Codes." The proposed model incorporates three feature representation methods, including information retrieval (IR), query quality (QQ), and distance. It utilizes a cascade strategy to leverage ensemble learning and deep learning effectively.
Our approach contains three steps of training:
1 Data Preprocessing
2 Feature Representation of Links
3 Cascade Forest
The article has been submitted to SMC2023 and currently shows the code for the first and second steps. The remaining parts will be published after the paper is accepted.