基于超图卷积神经网络的多行为感知服务推荐方法

陆佳炜,李端倪,王策策,徐俊,肖刚

Multi-behavior aware service recommendation based on hypergraph graph convolution neural network

Jia-wei LU,Duan-ni LI,Ce-ce WANG,Jun XU,Gang XIAO

表 2 MBSRHCNN模型与基线方法的对比实验结果

Tab.2 Accuracy and correlation comparison results of MBSRHCNN model and baseline method

模型 稀疏数据集 稠密数据集 HR@5 NDCG@5 HR@10 NDCG@10 MRR HR@5 NDCG@5 HR@10 NDCG@10 MRR AFMRec 0.381 0.358 0.384 0.352 0.372 0.412 0.407 0.428 0.404 0.383 WR-MSN 0.335 0.312 0.321 0.303 0.314 0.465 0.452 0.442 0.431 0.428 DHCN 0.342 0.303 0.352 0.323 0.328 0.383 0.336 0.371 0.342 0.334 HyperRec 0.481 0.474 0.513 0.489 0.452 0.512 0.498 0.527 0.515 0.482 MBHT 0.482 0.436 0.509 0.462 0.474 0.528 0.477 0.542 0.496 0.506 MB-GCN 0.387 0.322 0.419 0.349 0.364 0.363 0.311 0.392 0.327 0.339 MBSRHGNN 0.504 0.489 0.547 0.507 0.497 0.572 0.542 0.596 0.556 0.541 IMP/% 4.5 3.2 4.8 3.8 4.9 8.3 8.8 5.6 7.9 7.8