Table 5 
Evaluation results on the Douban corpus. Notations have the same meaning as those in Table 4. On R10@5, only SAN significantly outperforms baseline methods.
 MAPMRRP@1R10@1R10@2R10@5
TF-IDF 0.331 0.359 0.180 0.096 0.172 0.405 
RNN 0.390 0.422 0.208 0.118 0.223 0.589 
CNN 0.417 0.440 0.226 0.121 0.252 0.647 
LSTM 0.485 0.527 0.320 0.187 0.343 0.720 
BiLSTM 0.479 0.514 0.313 0.184 0.330 0.716 
  
Multi-View 0.505 0.543 0.342 0.202 0.350 0.729 
DL2R 0.488 0.527 0.330 0.193 0.342 0.705 
  
MV-LSTM 0.498 0.538 0.348 0.202 0.351 0.710 
Match-LSTM 0.500 0.537 0.345 0.202 0.348 0.720 
Attentive-LSTM 0.495 0.523 0.331 0.192 0.328 0.718 
SCNsingle 0.506 0.543 0.349 0.203 0.351 0.709 
SANsingle 0.508 0.547 0.352 0.206 0.353 0.720 
  
SCNlast 0.526 0.571 0.393 0.236 0.387 0.729 
SCNstatic 0.523 0.572 0.387 0.228 0.387 0.734 
SCNdynamic 0.529 0.569 0.397 0.233 0.396 0.724 
  
SANlast 0.536 0.581 0.393 0.236 0.404 0.761 
SANstatic 0.532 0.575 0.387 0.228 0.393 0.736 
SANdynamic 0.534 0.577 0.391 0.230 0.393 0.742 
 MAPMRRP@1R10@1R10@2R10@5
TF-IDF 0.331 0.359 0.180 0.096 0.172 0.405 
RNN 0.390 0.422 0.208 0.118 0.223 0.589 
CNN 0.417 0.440 0.226 0.121 0.252 0.647 
LSTM 0.485 0.527 0.320 0.187 0.343 0.720 
BiLSTM 0.479 0.514 0.313 0.184 0.330 0.716 
  
Multi-View 0.505 0.543 0.342 0.202 0.350 0.729 
DL2R 0.488 0.527 0.330 0.193 0.342 0.705 
  
MV-LSTM 0.498 0.538 0.348 0.202 0.351 0.710 
Match-LSTM 0.500 0.537 0.345 0.202 0.348 0.720 
Attentive-LSTM 0.495 0.523 0.331 0.192 0.328 0.718 
SCNsingle 0.506 0.543 0.349 0.203 0.351 0.709 
SANsingle 0.508 0.547 0.352 0.206 0.353 0.720 
  
SCNlast 0.526 0.571 0.393 0.236 0.387 0.729 
SCNstatic 0.523 0.572 0.387 0.228 0.387 0.734 
SCNdynamic 0.529 0.569 0.397 0.233 0.396 0.724 
  
SANlast 0.536 0.581 0.393 0.236 0.404 0.761 
SANstatic 0.532 0.575 0.387 0.228 0.393 0.736 
SANdynamic 0.534 0.577 0.391 0.230 0.393 0.742 
View Large
Close Modal
This Feature Is Available To Subscribers Only

or Create an Account

Close Modal
Close Modal
This site uses cookies. By continuing to use our website, you are agreeing to our privacy policy. No content on this site may be used to train artificial intelligence systems without permission in writing from the MIT Press.
Accept