NACCARATO ALESSIA
Objectives
The course aims to introduce the fundamental principles of statistical learning for data analysis, with particular attention to regression and classification problems in supervised settings. By the end of the course, students will be able to understand the main statistical models for predictive analysis; evaluate and select models using statistical criteria and validation methods; interpret model results and assess their performance; use R (and RStudio) software to implement basic statistical learning analyses. The course emphasizes an applied approach, with practical examples based on real data.Risultati Apprendimento Attesi
By the end of the course, students will have acquired basic knowledge of the main concepts and methods of Statistical Learning and will be able to apply introductory statistical learning techniques to the analysis of real-world data and to prediction problems.In particular, students will be able to:
By the end of the course, students will have acquired basic knowledge of the main concepts and methods of Statistical Learning and will be able to apply introductory statistical learning techniques to the analysis of real-world data and to prediction problems.
In particular, students will be able to:
- understand the fundamental concepts of Statistical Learning and distinguish between supervised and unsupervised learning;
- recognize and distinguish between regression and classification problems;
- understand the concepts of predictive accuracy, prediction error, overfitting, and underfitting;
- understand the bias–variance trade-off and its role in model selection;
- apply and interpret basic linear regression models and fundamental classification methods;
- understand and apply basic model selection and regularization techniques;
- understand and use the main decision-tree-based methods;
- understand the basic principles of unsupervised learning techniques, with particular emphasis on dimensionality reduction and clustering;
- assess and compare the predictive performance of different models using appropriate error measures and validation techniques, including cross-validation;
- apply the methods covered in the course to simple real-world datasets using appropriate statistical software;
- critically interpret the results of the analyses and communicate them using appropriate statistical terminology.
Programme
The course will cover the following topics:- Introduction to the main models of statistical learning;
- Prediction and classification problems: Recalls on linear regression and the main methods of unsupervised classification;
- Supervised classification: K-Nearest-Neighbours;
- Misclassification error;
- Resampling methods: cross validation and bootstrap;
- Decision tree-based methods: regression trees, classification trees, bagging, random forests, boosting.
- Introduction to semi-supervised classification methods;
- Use of the statistical environment R
Prerequisiti
Successful completion of the Statistics and Statistics for Big Data exams. Basic knowledge of linear algebra, inferential statistics, and linear models is also required.Core Documentation
James, G., Witten, D., Hastie, T., Tibshirani, R. (2021). An Introduction to Statistical Learning: with Applications in R (2nd ed.). Springer.Attendance
Attendance is not mandatory, but is strongly recommended.Type of evaluation
Oral examination. Two written midterm tests will be held during the course (the first approximately halfway through the course and the second towards the end of the course). Registration for the midterm tests must be completed exclusively through GOMP. The midterm tests are open to all students, whether attending or non-attending.