This is a sentimental analysis project that aims to provide a better insight on customers' satisfaction based on comments gathered (scrapped) from social media using google's Bert classification model. We focused on the main 3 spoken dialects in tunisia: French, English and tunisian. So by the end of this work we should have three classifiers, one for each dialect.
The execution requires having all files under the provided folder named 'files' placed under users/maily or users/'username'. we've devided the code into 3 notebooks, one for each dialect: final_bert_transfer_learning-english.ipynb, final_bert_transfer_learning-french.ipynb and final_bert_transfer_learning-tunisien.ipynb.
In order to execute the "final_bert_transfer_learning_french" notebook you need to create the following hierarchy:
"train" folder
"test" folder
"pos" and "neg" folders under "train"
"pos" and "neg" folders under "test"
DataEnoding.ipynb and answersGenerator.ipynb are related to the next part for which we needed to have our dataset restructured into a data warehouse to later feed it to PowerBI for analytical purposes.
This is a group project and all the following members have equally collaborated in the achievement of this work:
-Yessine Khanfir https://www.linkedin.com/in/yessine-khanfir-b5b509177/
-Sarrah Ferchichi https://www.linkedin.com/in/sarra-ferchichi-822a491ba/
-Rawia Khechini https://www.linkedin.com/in/rawia-khchini-11536a133/
-Fedi Baccouche https://www.linkedin.com/in/fedi-baccouche-283b361b8/
-Mayssa Zaouali https://www.linkedin.com/in/mayssa-zaouali-9288421bb/
-Wejdene Ben Jeddou https://www.linkedin.com/in/wejdene-benjeddou-154221182/
-Bennacef Mohammed Yassine https://www.linkedin.com/in/bennacef-mohamedyassine-732b4b202/