The files contain the following information:
-
DT.ipynb: in the first part, there are experiments made on monks 1, 2, 3 with random forests, using the grid search. In the second part, there are experiments made on monks 1, 2, 3 with random forests. Both models are been implemented using scikit library;
-
DT_ml_cup22.ipynb: experiments made for the ML Cup using decision trees and random forests. In the first part, there are experiments made with a decision tree model, using the grid search, while in the second part there are experiments with random forests. Both models are been implemented using scikit library;
-
keras_RandomSearch.ipynb: experiments on monks 1, 2, 3 using neural networks implemented with Keras library. In this notebook is used the random search to look for the best combination of hyperparameters;
-
KNN.ipynb: experiments on monks 1, 2, 3 using knn models implemented with scikit library;
-
ml_cup22_data_splitter.ipynb: creation of an internal design set and test set using the ML-CUP22-TR.csv file;
-
ml_cup22_keras_RS.ipynb: experiments made for the ML Cup using neural networks implemented with Keras library. In this notebook is used the random search to look for the best combination of hyperparameters;
-
ml_cup22_model_assessment.ipynb: model assessment initially made individually on each model and then using the ensembling method;
-
ml_cup22_pyTorch_GS.ipynb: experiments made for the ML Cup using neural networks implemented with PyTorch library;
-
ML_cup_KNR.ipynb: experiments made for the ML Cup using the k-nearest regressor implemented with scikit library;
-
ML_cup_KRR.ipynb: experiments made for the ML Cup using the Kernel Ridge Regressor implemented with scikit library;
-
ML_cup_SVM.ipynb: experiments made for the ML Cup using the k-nearest regressor implemented with scikit library;
-
pytorch_GridSearch.ipynb: experiments on monks 1, 2, 3 using neural networks implemented with PyTorch library;
-
SVM.ipynb: experiments on monks 1, 2, 3 using SVM implemented with scikit library;
-
final_res.md: text file containing Mean Euclidean Distance and other measures for each model used for the ensembling.