An End-to-End Review-Based Aspect-Level Neural Model for Sequential Recommendation

Table 4

Comparison of the model variants for the Toys, Video, Beauty, and Book datasets.

Architecture	Toys	Video	Beauty	Book
No SP	0.934	1.023	1.201	0.988
No AP and no SP	1.035	1.362	1.385	1.156
No PT	0.956	1.122	1.212	1.026
No SA	0.962	1.343	1.320	1.089
No PE	1.023	1.025	1.271	1.088
No SM	1.123	1.401	1.482	1.293
Default	0.920	0.985	1.198	0.942

The worst and second-worst results are highlighted in bold and underlined, respectively.