Hi all,
since the deadline is approaching, we want to make you aware of some final points:
1) The deadlines for both subtasks have been extended to 31.01. 23:59:59 UTC.
2) Make sure to include at least the fields 'id', 'sentiment score' and 'cashtag' in your submissions.
3) You can still request the
complete training and test data for subtask 1 (with messages) if you didn't do so yet (email to
semeval...@mail.com).
4) All of you have 2 submissions per subtask available. We would recommend you to submit twice since we take only the best of your two submissions into consideration for the final ranking.
5) Since we are interested in your approach, we would appreciate information regarding the following:
- Approach adopted:
- Pre-processing (e.g., tokenization, stemming, lemmatization, POS tagging, emoticon removal, etc.)
- Techniques used (e.g., hybrid, machine learning based, rule based, ontology-based, etc.)
- External sources, data sets and/or lexicons used (if any)
- Tools used (if off the shelf or ones implemented for the scope of the task/s)
- Why did you choose this approach and if it is: (i) multilingual/cross-lingual, (ii) domain dependent/independent
- Issues encountered (if any) and how they were tackled and potentially solved
Please send your information to semeval...@mail.com and use as subject "Subtask
X - Approach of team
TEAMNAME". Please replace
X and
TEAMNAME with the proper information.
Best,
Tobias