AI4D iCompass Social Media Sentiment Analysis for Tunisian Arabizi
Helping Tunisia
$2 000 USD
Completed (almost 5 years ago)
Natural Language Processing
Classification
Sentiment Analysis
902 joined
309 active
Start
Nov 20, 20
Close
Mar 28, 21
Reveal
Mar 28, 21
Starter notebook
Data · 23 Feb 2021, 21:30 · 2
Here's a simple python notebook with some ML models (classification) applied on the social media tunisian texts that I advice everyone who hasn't submitted yet to visit, try it and tweak parameters of the models to get better results.
https://github.com/Houssem96/Social-Media-Sentiment-Analysis-for-Tunisian-Arabizi/blob/main/Tunisian_sentiment_analysis_starter_notebook.ipynb
Thanks... But you can't go far with this, i think there is much about preprocessing or even a better model for this challenge... or a methodology.
All right, it's only a starting notebook; Of course there's a lot of text preprocessing (eleminating stop words, stemming, correction...) and also there's an unbalanced class which is the 0 class label to handle.