Zelko, Andreea Cristina (2025) Towards Identifying Software Sustainability within Practitioners’ Cloud Computing Q&A Discussions. Master's Thesis / Essay, Computing Science.
|
Text
mCS2025ZelkoAC.pdf Download (6MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
Toestemming.pdf Restricted to Registered users only Download (159kB) |
Abstract
Sustainability has become an increasingly important consideration in software architecture and development. It is crucial to understand how practitioners think and talk about this topic. This study takes a complementary approach to existing work by examining how sustainability is addressed in developer conversations on public platforms. Specifically, StackOverflow posts are analyzed to gain insight into how developers engage with software sustainability in their day-to-day discussions. To do this, a qualitative content analysis codebook grounded in theoretical models of sustainability was developed iteratively. The codebook was then used to manually annotate a dataset of 592 StackOverflow posts. The annotation task was also an iterative process, which was distributed among four annotators. Throughout the annotation process, the inter-rater agreement was assessed using Krippendorff's Alpha, and the average score was 0.42. This study further investigated the feasibility of automating the annotation process using large language models (LLMs). Several models were trained, using both prompt-engineering and fine-tuning strategies. The models were evaluated on the tasks of binary sustainability detection and multi-label classification of specific sustainability dimensions. The best model shows promising results, achieving a 0.83 macro F1-score for sustainability classification.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Andrikopoulos, V. and Ahmadisakha, S. |
| Degree programme: | Computing Science |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 07 Jul 2025 11:56 |
| Last Modified: | 07 Jul 2025 11:56 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/35900 |
Actions (login required)
 |
View Item |