Multi-channel Convolutional Neural Network Feature Extraction for Session Based Recommendation

Table 5

The performance of various K (P5 means P@K with , M5 means MRR@K with , etc.).

Datasets	Yoochoose 1/64						Diginetica
Measures	P5	M5	P10	M10	P20	M20	P5	M5	P10	M10	P20	M20
NARM	44.34	26.21	57.50	27.97	68.32	28.63	27.95	15.67	39.89	17.26	49.70	16.17
STAMP	45.69	27.26	58.07	28.92	68.74	29.67	20.41	10.87	30.46	12.20	45.64	14.32
SR-GNN	47.26	28.26	60.09	30.69	70.57	30.94	26.64	15.10	37.91	16.61	50.73	17.59
RMMCNN	47.61	28.87	60.32	30.71	70.94	31.46	26.83	15.90	36.64	17.93	51.82	18.19