This project has been developed as part of the exam of the Machine Learning class @ UniPi

Objectives

The main goals of the project are

• the comparison and evaluation of different ML models implemented by different frameworks;
• the evaluation of the usability of the frameworks;
• the evaluation of the accuracy and the performances of the models;
• the identification of the best-performing model to compete in an internal competition among students.

Tasks

The models should be applied and tested on two different tasks:

• The Monk's Problem to start with a simple classification task on which to apply the ML techniques learned during the course;
• A regression task known as The Cup: the lecturer provided us with a test set for which to estimate the target values by means of our best-performing model.

Results

We tested and compared a long serie of different models from different Python frameworks, here you can find our NN implementations. You can find more informations into the project report.