BMAM: complete the missing POI in the incomplete trajectory via mask and bidirectional attention model

EURASIP Journal on Wireless Communications and Networking

Table 2 Performance of BMAM and baseline on the dataset

Methods	Metrics	STRNN	PACE	Bi-STDDP	SASRec	BERT4Rec	BMAM	Improv (%)
NYC	Recall@5	0.3225	0.3226	0.3476	0.5411	0.6005	0.6191	3.1
	Recall@10	0.4055	0.4033	0.4176	0.6417	0.6606	0.6756	2.3
	F1-score@5	0.1075	0.1075	0.1159	0.1804	0.2002	0.2063	3.0
	F1-score@10	0.0737	0.0733	0.0759	0.1167	0.1201	0.1228	2.2
TKY	Recall@5	0.3665	0.3605	0.4107	0.6249	0.6631	0.6852	3.3
	Recall@10	0.4492	0.4469	0.4838	0.7021	0.7256	0.7386	1.8
	F1-score@5	0.1222	0.1202	0.1369	0.2083	0.2210	0.2284	3.3
	F1-score@10	0.0817	0.0813	0.0880	0.1277	0.1319	0.1343	1.8