Week 7&8

Analyses for Previous Works

Posted by Stuart Chen on July 19, 2019

The Statistics

The Scores For The Answers

This is the table for GERBIL evaluation.

Method BLEU score Micro F1 Micro Precision Micro Recall Macro F1 Macro Precision Macro Recall
DBNQA 64.55 0.0072 0.0274 0.0041 0.0093 0.0093 0.0093
QALD7 80 0.0072 0.0282 0.0041 0.0093 0.0093 0.0093
dependency parsing + Spotlight 80 0.0037 0.0179 0.0021 0.0047 0.0047 0.0047
dependency parsing + Spotlight + GloVe 80 0.0072 0.027 0.0041 0.0093 0.0093 0.0093

The Scores For The Answer Type

This table shows the statistics for answer types evaluation.

Method BLEU score Micro F1 Micro Precision Micro Recall Macro F1 Macro Precision Macro Recall
DBNQA 64.55 0.6587 0.9167 0.514 0.5163 0.5163 0.5163
QALD7 80 0.6587 0.9167 0.514 0.5163 0.5163 0.5163
dependency parsing + Spotlight 80 0.6587 0.9167 0.514 0.5163 0.5163 0.5163
dependency parsing + Spotlight + GloVe 80 0.6587 0.9167 0.514 0.5163 0.5163 0.5163