ML Laboratory Repo (Sapienza University of Rome)

This repository contains the laboratory exercises with discussions of the Machine Learning course (2023/24) at the Master's degree in Computer Science at Sapienza University of Rome. For completeness purposes, you'll also find last year's (2022/23) laboratory lessons. This is to show you that the course is in continuous development and evolution.

The coding environment is Google Colab so that students don't have to configure a designated environment with specific Python packages.

The syllabus of the laboratory courses is:

Data pre-processing + Simple ML Models (lab 1)

Data feature pre-processing

Data cleaning - missing data.

Encoding - pitfalls of encoding categorical data, one-hot encodings

Simple ML Models - Decision Trees, Random Forests, XGBoost

XGBoost details - hyperparameters, optimization + overfitting

Naive Bayes, Linear Regression, and SVMs (lab 2)

Naive Bayes, Linear Regression, and SVMs

Bayes classification, Bayes's Theorem

Gaussian and Multinomial Naive Bayes

Simple Linear Regression + Basis Function for nonlinear feature relationships

Ridge and Lasso regularization

Simple insights on uncertainty

Linear vs nonlinear separation hyperplanes

Kernel trick - linear, polynomial, and radial basis function (RBF) kernel

Soft margins of SVMs