This project explores the application of decision trees and random forest classifiers to predict the type or country of origin of coffee and wine using a dataset containing relevant features. We implemented both models, leveraging the interpretability of decision trees and the robustness of the random forest ensemble approach.
To enhance model performance, we conducted hyperparameter tuning using grid search, aiming to identify the optimal configurations for each classifier. Finally, we compared the performance of our custom implementations against the standard versions provided by scikit-learn, assessing their effectiveness and accuracy.
The full report detailing the methodology, implementation, and results can be found in the repository under the file report.pdf