Istudor, Andrei-Stefan (2024) Applying Graph Learning For Technical Debt Detection. Bachelor's Thesis, Computing Science.
|
Text
bCS2024IstudorAS.pdf Download (2MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming.pdf Restricted to Registered users only Download (143kB) |
Abstract
In the domain of software engineering, the increasing presence of Technical Debt (TD) poses a significant challenge for the quality and long-term maintenance of modern software systems. Technical Debt refers to the extra cost and effort required, due to early sub-optimal decisions in software development. This research project seeks to apply graph learning techniques in order to identify Technical Debt within software projects. We developed a pipeline called ‘Debtective’ for preprocessing code samples and for model training. This allowed us to investigate the effectiveness of different model architectures in detecting Technical Debt from a chosen dataset. We selected and transformed code samples for two SonarQube code smells into Code Property Graphs (CPGs), and we used them in training and evaluating a series of models that we created. The results show that some models perform better for a code smell compared to the other, and some models perform great generally, for both code smells. Overall, the results of our study confirm the effectiveness of graph-based models in detecting Technical Debt (TD), and highlight its potential for future TD detection research. Through this proposed approach, we hope to enhance the identification of TD and offer new methodologies for its detection, potentially benefiting both the industry practices and the academic research in this area.
| Item Type: | Thesis (Bachelor's Thesis) |
|---|---|
| Supervisor name: | Feitosa, D. and Maarleveld, J. |
| Degree programme: | Computing Science |
| Thesis type: | Bachelor's Thesis |
| Language: | English |
| Date Deposited: | 01 Aug 2024 10:14 |
| Last Modified: | 01 Aug 2024 10:14 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/33790 |
Actions (login required)
 |
View Item |