Leveraging Relational Concept Analysis for Automated Feature Location in Software Product Lines
summary
Proposes a concrete pipeline that takes two formal contexts (features-per-product and artefacts-per-product), builds FCA AOC-posets, links them with Relational Concept Analysis into a relational AOC-poset, extracts core and interaction traces, converts them to boolean formulas and applies four-step simplification. Evaluated on five Java case studies with 100% artefact reconstruction and a 67% average trace reduction (reported 31–85% per case), claims automated, fine-grained feature and interaction location that reproduces original artefact sets.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:20Feature Location
- 2:15General Concept
- 4:44Feature Location Process
- 5:00Concept initialization
- 5:37Concept analysis
- 6:58Apparition context
- 7:20Positive feature
- 7:55Negative feature
- 8:35Rate concept
- 9:52Experiment
- 10:42Results
- 11:21thresh reduction
- 11:31threat of validity
- 12:28Related work
- 13:47Conclusion
- 14:32Future work
related talks
30mAn Extensible Feature-Oriented Approach for Fine-Grained Code Quality Analysis
15mCorrectness-by-Construction for Feature-Oriented Software Product Lines
40mRelational Reasoning in Object-based Programs
23mUnleashing the Power of Implicit Feedback in Software Product Lines: Benefits ahe...
36mStatic Analysis of APL for Tooling and Compliance
45mWhere AI Meets Code
44mAI Architecture Thinking: From Idea to Impact
30mAutomatic Inference of Relational Object Invariants