The Construction and Information Supplementation of the Knowledge Graph Applied to Software Changes

The knowledge graph is increasingly being utilized to address issues such as defect detection and test case generation in software engineering. However, constructing knowledge graphs presents challenges in contexts where software changes frequently. To address this, this paper proposes a method for constructing and supplementing information in knowledge graphs for changed software. Firstly, define and extract the triple elements of the software knowledge graph by using the code and related test logs. Then, supplement the historical data related to the changes based on the keyword search pattern, and construct the serialized vector model of entity attributes. Finally, build the knowledge graph of the changed software that integrates the execution logic. The proposed method was applied to $\mathbf{1 0}$ different scale datasets and compared with other three algorithms. Experimental results show that the proposed method can significantly improve the construction efficiency of the change software knowledge graph while greatly reducing the parsing cost of multi-source heterogeneous data.