Research on Domain-Specific Knowledge Graph Based on the RoBERTa-wwm-ext Pretraining Model

Table 8

Named entity model recognition results.

Label	Precision (%)	Recall (%)	F1	Number statistics
NAT (nation)	85.87	93.83	89.68	8453
PLA (plane)	79.05	76.12	77.55	1532
VES (vessel)	66.71	69.49	68.07	775
MIS (missile)	61.60	50.10	55.26	388
WAT (waters)	58.33	69.72	63.52	300
AIR (aircraft carrier)	63.64	52.34	57.44	88
BAS (military base)	50.51	48.54	49.50	99
RAD (radar)	72.46	37.88	49.75	69
ISL (island)	60.92	64.63	62.72	87
Overall	81.11	85.13	83.07	11722