Clara Marie Lüders has successfully defended her Doctoral Thesis at the University of Hamburg with excellent results. Congrats Dr. Clara!
📖 Thesis: Mining and Understanding Issue Links Towards a Better Management
Clara extensively studied issue dependencies in issue trackers, briefly speaking along three main lines: exploring, predicting, and designing:
✅ While most research has focussed on duplicate links so far, Clara showed that other links are very popular and effortful to manage. Her analysis shows the heterogeneity of issue linking and variety of link types as well as issue linking patterns in (open source) practice.
✅ Clara showed that top state-of-the-art duplicate prediction models are rather unreliable if exposed to realistic datasets with different link types (beyond duplication). She experimented with multiple Machine Learning techniques (including Transformers) for predicting typed issue links with top accuracy: researching, a.o. the impact of data quality (i.e. of links and issues) on the prediction accuracy.
✅ Clara designed and evaluated a software framework for link visualization, link recommendation, and linking analytics. The framework is based on and integrated into hashtag#Jira. We are considering sharing it as a plugin in their store.
These are the main papers published as a part of her PhD journey.
- Automated Detection of Typed Links in Issue Trackers
- OpenReq Issue Link Map: A Tool to Visualize Issue Links in Jira
- Beyond Duplicates: Towards Understanding and Predicting Link Types in Issue Tracking Systems
- An Alternative Issue Tracking Dataset of Public Jira Repositories. (might be useful for many working in this area).
A warm congratulation and expect more to come from Clara!
