Master Degree (Artificial Intelligence curriculum)
ML course 654AA, Academic Year: 2021/2022
Date: 04/01/2022
Type of project: A
Teacher: Alessio Micheli.
Project implementation for Machine Learning exam, MD Course in Computer Science, University of Pisa.
This report describes the Python implementation of an Artificial Neural Network, capable to deal with both classification and regression tasks, and shows its results on Monk and ML-CUP21 problems. The final model for the CUP problem was selected through a grid search over hundreds of hyper-parametric configurations using the K-Fold cross-validation strategy and then tested on an internal test set with Hold Out
| File | Description |
|---|---|
| main_monk.py | Our best model for MONKs' problems |
| main_cup.py | Our best model for CUP's problem |
| model_selection_monk.py | The starting point of model selection for MONKs' problems |
| model_selection_cup.py | The starting point of model selection or CUP's problem |
| model_selection_cup_distributed.py | The starting point of distribuited model selection or CUP's problem (Note: it requires a database and its initialization) |
| AIAIAI_ML-CUP21-TS.csv | Our Results for the Blind TS |
| report.pdf | Our report |