Test data versus Training data

[Caroline Brun]

Dear organizers

It seems  that most of the sentences (~95 %)  present in TrialData_SubtaskA_Test.csv,  are included in the full training set, Training_Full_V1.1.csv, and that
 the few sentences left are just slight variants (corrected) of sentences present in the Train set.

Is this normal ? During the trial period,  do we have indeed  to score TrialData_SubtaskA_Test.csv, i.e. a subsample of the training set ?

Thanks to clarify this point.

Best regards,

Caroline

[Sapna Negi]

Dear Caroline,

No, this wasn’t supposed to be the case. 

Thanks for pointing this out. We may have mixed up the files, we will rectify this asap.

Regards

--
You received this message because you are subscribed to the Google Groups "SemEval 2019: Task 9" group.
To unsubscribe from this group and stop receiving emails from it, send an email to semeval-2019-ta...@googlegroups.com.
To post to this group, send email to semeval-2...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/semeval-2019-task-9/085d05d8-da5f-4683-9a01-479873ed5763%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Sapna Negi]

Dear participants

Further to the reporting of trial test data being present within the training set,  we have updated the GitHub repo with the new version of train dataset Training_Full_V1.2.csv where the test instances are removed.

https://github.com/Semeval2019Task9/Subtask-A

Apologies for any inconvenience caused. We also request the participants to re-submit their new results on Codalab, using the updated train set.

Also, we will be releasing the final version of full train set in a few days with additional train sentences.

Regards

Sapna