BTK Academy in collaboration with Google & Girişimcilik Vakfı have created a datathon (Datahon2023) in the form of a competition where groups are expected to build models on the given training set, perform an 8-class classification, and obtain the best performance on the unlabeled test set (where the labels are only available within the Kaggle's submission system).
Our group, named Troia, in honor of our city Canakkale (formerly known as the ancient city Troia) and our university Canakkale Onsekiz Mart University, have obtained the highest score together with nine other groups in this Kaggle competition.
* image created by DALL.E2
In the jupyter notebook file you can see in this repository, you can see our code that helped us reach this result.
For both members of this team to code together simultaneously, we have used Deepnote to create the jupyter notebook. For an unknown reason, when we upload our notebook to Github, after some cells, Markdowns and code cells have merged and destroyed the visuals. Thus, in order to view and try our code, and examine our results, please feel free to use the following alternative links:
Google: https://drive.google.com/file/d/10FRb6O9jU1d39g_1PnNhFvqYi_iObxyi/view?usp=sharing
Kaggle: https://www.kaggle.com/code/ulyabayram/datathon2023-troia-notebook
Deepnote: https://tinyurl.com/datathon2023Deepnote