Table 4 Basic information about the datasets.

From: LLM-based intelligent Q&A system for railway locomotive maintenance standardization

Dataset type

Precision/%

Recall/%

F1/%

Runtime/s

Original dataset

93.65

93.28

93.46

6338.74

Expanded dataset

93.64

91.78

92.70

10953.33

Clustered dataset

90.66

88.81

89.73

4823.33