Cross-level requirements tracing based on large language models
Publication Type
Journal Article
Publication Date
7-2025
Abstract
Cross-level requirements traceability, linking high-level requirements (HLRs) and low-level requirements (LLRs), is essential for maintaining relationships and consistency in software development. However, the manual creation of requirements links necessitates a profound understanding of the project and entails a complex and laborious process. Existing machine learning and deep learning methods often fail to fully understand semantic information, leading to low accuracy and unstable performance. This paper presents the first approach for cross-level requirements tracing based on large language models (LLMs) and introduces a data augmentation strategy (such as synonym replacement, machine translation, and noise introduction) to enhance model robustness. We compare three fine-tuning strategies—LoRA, P-Tuning, and Prompt-Tuning—on different scales of LLaMA models (1.1B, 7B, and 13B). The fine-tuned LLMs exhibit superior performance across various datasets, including six single-project datasets, three cross-project datasets within the same domain, and one cross-domain dataset. Experimental results show that fine-tuned LLMs outperform traditional information retrieval, machine learning, and deep learning methods on various datasets. Furthermore, we compare the performance of GPT and DeepSeek LLMs under different prompt templates, revealing their high sensitivity to prompt design and relatively poor result stability. Our approach achieves superior performance, outperforming GPT-4o and DeepSeek-r1 by 16.27% and 16.8% in F-measure on cross-domain datasets. Compared to the baseline method that relies on prompt engineering, it achieves a maximum improvement of 13.8%.
Keywords
Feature Extraction, Semantics, Deep Learning, Information Retrieval, Data Augmentation, Software, Vectors, Training, Large Language Models, Accuracy, Requirements Tracing, Large Language Models, Fine Tuning, Data Augmentation, Software Requirements, Language Model, Large Language Models, Machine Learning, Deep Learning, Superior Performance, Machine Learning Methods, Data Augmentation, Poor Stability, Information Retrieval, Baseline Methods, Traditional Machine Learning, Machine Translation, Traditional Machine Learning Methods, Fine Tuning Strategy, Model Performance, Positive Samples, Machine Learning Models, Deep Learning Models, Precision And Recall, Words In Sentences, Data Augmentation Techniques, High Recall, Original Text, Recall Rate, Vector Space Model, CNN Model, Video Summarization, Jensen Shannon Divergence, Fine Tuning Method, Latent Dirichlet Allocation
Discipline
Artificial Intelligence and Robotics
Research Areas
Intelligent Systems and Optimization
Publication
IEEE Transactions on Software Engineering
Volume
51
Issue
7
First Page
2044
Last Page
2066
ISSN
0098-5589
Identifier
10.1109/TSE.2025.3572094
Publisher
Institute of Electrical and Electronics Engineers
Citation
GE, Chuyan; WANG, Tiantian; YANG, Xiaotian; and TREUDE, Christoph.
Cross-level requirements tracing based on large language models. (2025). IEEE Transactions on Software Engineering. 51, (7), 2044-2066.
Available at: https://ink.library.smu.edu.sg/sis_research/10806