软件导刊2025，Vol.24Issue(7)：73-81,9.DOI:10.11907/rjdk.241492
基于特征融合与复制增强的源代码摘要模型
A Source Code Summary Model Based on Feature Fusion and Copy Enhancement
摘要
Abstract
To solve the problem of some specific information loss of CodeBERT model in the task of generating abstracts from source code,a feature fusion and copy-enhanced source code abstract model CodeRoBERT is proposed.The feature fusion layer assigns different weights to features by the attention mechanism and adds up the weights.More attention is paid to the low frequency tokens of particular categories by the replication mechanism.The experimental results of Ruby,Java,JS and PHP source code abstracts generated on the CodeSearchNet open data-set show that the BLEU-4,ROUGE-1,ROUGE-2 and ROUGE-L performance of the CodeRoBERT model outperformed the contrast models.关键词
复制机制/特征融合/源代码摘要/注意力机制Key words
copy mechanism/feature fusion/source code summarization/attention mechanism分类
信息技术与安全科学引用本文复制引用
梁岳,董东..基于特征融合与复制增强的源代码摘要模型[J].软件导刊,2025,24(7):73-81,9.基金项目
河北省高等学校人文社会科学研究项目教育发展专项(WTZX202421) （WTZX202421）
河北师范大学人文社会科学研究基金项目(S23JX003) （S23JX003）
