Research Article

Session Recommendation Model Based on Context-Aware and Gated Graph Neural Networks

Table 3

Comparison of the experimental results of CA-GGNN with TAGNN and DGTN (%).

MethodYoochoose1/64Yoochoose1/4Diginetica
P@20MRR@20P@20MRR@20P@20MRR@20

TAGNN70.1730.4671.0530.9650.0217.18
DGTN70.2131.0871.3831.3750.3517.33
CA-GGNN70.8431.8372.9332.9151.1218.48
Improve0.630.751.551.180.771.15

Bold shows the experimental result of the model proposed in this paper.