An End-to-End Review-Based Aspect-Level Neural Model for Sequential Recommendation
Table 3
Performance comparison in terms of MSE.
Dataset
PMF
RBLT
CMLE
DeepCoNN
D-Attn
TransNet
TARMF
MPCN
ANR
CARP
RANM
Amazon
1.367
1.325
1.105
1.192
1.218
1.111
1.285
1.352
1.129
1.158
0.983
Apps
1.601
1.582
1.562
1.552
1.337
1.625
1.375
1.402
1.412
1.325
1.201
Automotive
1.325
1.158
1.259
1.426
1.313
1.205
1.024
0.967
1.151
0.978
0.945
Baby
1.385
1.368
1.258
1.469
1.309
1.421
1.306
1.313
1.258
1.111
1.009
Beauty
1.455
1.572
1.596
1.578
1.389
1.458
1.398
1.385
1.299
1.255
1.198
Books
1.185
1.205
1.098
1.125
1.220
1.024
1.052
1.156
0.985
0.989
0.942
CDs
1.256
1.156
1.203
0.952
1.186
1.001
1.123
1.054
0.926
0.930
0.910
Cell phones
2.452
2.412
2.045
2.118
1.625
1.983
1.689
1.489
1.675
1.652
1.365
Clothing
1.528
1.569
1.452
1.425
1.384
1.356
1.386
1.124
1.255
1.023
1.005
Digital music
1.206
0.870
0.885
1.045
0.925
0.913
0.853
0.989
0.889
0.820
0.789
Electronics
1.856
1.985
1.825
1.657
1.356
1.658
1.698
1.364
1.156
1.256
1.106
Grocery
1.489
1.568
1.425
1.445
1.395
1.352
1.258
1.135
1.256
1.105
1.004
Health
1.725
1.689
1.652
1.562
1.336
1.425
1.325
1.245
1.348
1.289
1.204
Home
1.852
1.762
1.714
1.528
1.425
1.475
1.468
1.235
1.326
1.302
1.210
Kindle store
1.333
1.258
1.098
0.917
0.986
0.993
0.856
0.786
0.878
0.798
0.724
Movie
1.526
1.583
1.356
1.288
1.285
1.111
1.200
1.156
1.152
1.017
1.009
Music
1.398
0.815
0.817
1.175
1.302
0.798
0.943
0.846
1.052
0.773
0.705
Office
1.092
0.759
0.759
0.729
0.858
1.259
0.789
0.795
1.096
0.719
0.706
Patio
1.785
1.726
1.689
1.633
1.489
1.521
1.485
1.046
1.354
1.258
1.003
Pet supplies
1.658
1.628
1.521
1.595
1.446
1.485
1.460
1.385
1.352
1.294
1.213
Sport
1.428
1.596
1.324
1.301
1.275
1.215
1.115
1.134
1.103
1.059
1.005
Tools
1.566
0.983
1.020
1.487
1.225
1.003
1.169
1.082
1.256
0.960
0.952
Toys
1.396
1.425
1.305
1.236
1.106
1.112
1.213
0.985
1.093
1.007
0.920∗
Video
1.672
1.143
1.253
1.445
1.346
1.390
1.195
1.301
1.187
1.084
0.985
Yelp18
1.789
1.987
1.852
1.669
1.458
1.602
1.582
1.469
1.556
1.413
1.398
Average
1.533
1.445
1.363
1.382
1.288
1.3
1.25
1.159
1.206
1.103
1.02
The best and second-best results are highlighted in bold and underlined, respectively. A statistical significance test (using ∗ to show the result) is conducted by performing a pairwise Student’s t-test.