Session-Based Recommendation with GNN and Time-Aware Memory Network

Table 2

The performance of different methods over three datasets (%).

Method	Yoochoose1/64		Yoochoose1/4		RetailRocket
	P@20	MRR@20	P@20	MRR@20	P@20	MRR@20
Item-KNN	51.60	21.81	52.31	21.70	24.01	10.72
FPMC	45.62	15.01	—	—	30.21	25.73
GRU4REC	60.64	22.89	59.53	22.60	48.32	24.31
NARM	68.32	28.63	69.73	29.23	53.04	29.43
STAMP	68.74	29.67	70.44	30.00	53.94	28.49
SR-GNN	70.57	30.94	71.36	31.89	54.33	30.66
DPAN4Rec	71.04	30.50	71.52	30.89	55.63	31.17
SLIST	70.88	30.83	71.75	31.61	60.20	35.12
NISER+	71.27	31.61	71.80	31.77	61.41	36.50
CSRM	69.85	29.71	70.63	29.48	54.12	30.48
STAN	69.45	28.74	70.07	28.89	59.80	35.32
SR-GTM	71.82	32.53	72.07	32.57	63.64	37.41