This project consists in comparing different tools and models to solve a regression task, with given train and test sets.
The whole project has been written in Python 3.7.
Create a virtual environment , and install the dependecies:
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
Classification Task - MONK's Problem
We developed a little Neural Network, expoliting Keras library, to solve 3 (plus one with regularization) classification tasks. Results can be seen in Section 3.4 of the report.
We compared two models: Neural Networks (NN) and Support Vector Machines (SVM), exploiting 3 different libraries:
- Keras and PyTorch (for NN)
- scikit-learn (for SVM)
For both models we used a validation schema consisting into an exhaustive grid search and K-Fold Cross-validation tecnique for model selection and hyperparameters' tuning.
Please, read the report for a deeper description of our work.