|
Development and comparative evaluation of large language models for automated test case generation in banking software testing |
|---|---|
| รหัสดีโอไอ | |
| Creator | Hathairat Janwittaya |
| Title | Development and comparative evaluation of large language models for automated test case generation in banking software testing |
| Contributor | Ratthaslip Ranokphanuwat |
| Publisher | Mahasarakham University |
| Publication Year | 2569 |
| Journal Title | Journal of Science and Technology Mahasarakham University |
| Journal Vol. | 45 |
| Journal No. | 4 |
| Page no. | 446-462 |
| Keyword | Large language models, software testing, test case generation, test coverage metrics, vector embedding |
| URL Website | https://li01.tci-thaijo.org/index.php/scimsujournal |
| Website title | Journal of Science and Technology Mahasarakham University |
| ISSN | 1686-9664 (Print), 2586-9795(Online) |
| Abstract | This research develops and evaluates four Large Language Models for automated test case creation in banking systems. The methodology employs a dataset of 20,000 banking transaction test cases, utilizing LoRA fine-tuning with LangChain and ChromaDB vector database through RAG architecture. Results show Qwen3-8B achieved the lowest loss (0.1558) for learning accuracy, while Gemma-3-4b obtained superior similarity metrics and the highest Test Coverage (93%) with expert evaluation scores of 86.4%. The study concludes that LLMs significantly enhance testing efficiency and coverage. |
