基于Word2Vec和决策树的故障定位技术
Fault location technology based on Word2Vec and decision tree
王露露 1陈军华1
作者信息
- 1. 上海师范大学 信息与机电工程学院,上海 201418
- 折叠
摘要
利用Word2Vec方法对Java源代码进行深层语义编码,生成文件级和行级的语义向量,并将其用作输入数据来训练决策树模型,以实现精确的文件级别和行级别故障定位,优化故障检测过程,构建一个综合文件级别与行级别分析的高效故障定位框架.实验结果表明:该模型在各项目中的故障定位准确率均高于83%.
Abstract
Word2Vec technology was utilized to perform deep semantic encoding on Java source code,generating file-level and line-level semantic vectors.These vectors were used as input data to train the decision tree model,aiming to achieve precise file-level and line-level fault location and to optimize the fault detection process.An efficient fault localization framework was constructed by this method which integrated file-level and line-level analysis.The experimental results showed that the fault localization accuracy of the model in all projects was higher than 83%.
关键词
故障定位/语义表示/Word2Vec/决策树Key words
fault location/semantic representation/Word2Vec/decision tree引用本文复制引用
出版年
2024
NETL
NSTL
万方数据