AI4D iCompass Social Media Sentiment Analysis for Tunisian Arabizi
$2,000 USD
Can you classify sentiment in the Tunisian Arabizi dialect?
754 data scientists enrolled, 311 on the leaderboard
MediaClassificationPredictionUnstructuredNLP
Tunisia
20 November 2020—28 March 2021
129 days
I got disqualified, but here's my solution nonetheless.
published 8 Jun 2021, 17:40

I got the disqualified for alledegly "collaborating out of team". However if my score wasn't redacted from the leaderboard, I would've been the first still.

So I thought my solution would be of interest to some of you: https://github.com/maroxtn/tun-sentiment

Were you the one with a score over 0.9?

No no, that one was probably put by the organizers. Mine is 0.848 on the private leaderboard.

Great solution and thought process. Do you think the transliteration tools built would have been permitted even though it's based on external data. Thanks for sharing.

Thanks ! That I was not sure of, my thought process was that they may allow it because it iis my own personal dataset

I'm not so sure about that. All the same, very sophisticated solution. Probably the best way to have approached the challenge if it were an actual project rather than a competition with uncertain rules capable of impeding the quality. Awesome work!