The course aims to explore the fundamental concepts of statistical learning, introducing a wide range of tools and techniques such as linear and nonlinear models, decision trees, neural networks, and dimensionality reduction methods. The main objective is to provide both theoretical knowledge and practical skills in data manipulation and analysis, as well as in the evaluation and interpretation of model results with particular attention to supervised learning problems. Real-world problems will be addressed to understand the theoretical principles underlying learning algorithms and to learn their effective application using specific libraries of the statistical software R.
Curriculum
teacher profile teaching materials
Review of multiple linear regression models. Basis functions for regression: polynomials, piecewise polynomials, and splines. Truncated basis functions and selection of the number of knots.
Indirect estimation of test error: Mallows' Cp, AIC, and BIC. Direct estimation of test error: validation set, LOOCV (Leave-One-Out Cross-Validation), K-fold cross-validation, and bootstrap. Ridge regression and lasso regression.
Introduction to classification problems. Bayes classifier and Naive Bayes. k-Nearest Neighbors (KNN), confusion matrix, and derived performance metrics.
Introduction to regression and classification decision trees. Tree pruning. Ensemble methods: bagging, random forests, and boosting.
Programme
Introduction to statistical learning. Training error and test error: implications of model flexibility in regression. Decomposition of test error. Bias–variance trade-off.Review of multiple linear regression models. Basis functions for regression: polynomials, piecewise polynomials, and splines. Truncated basis functions and selection of the number of knots.
Indirect estimation of test error: Mallows' Cp, AIC, and BIC. Direct estimation of test error: validation set, LOOCV (Leave-One-Out Cross-Validation), K-fold cross-validation, and bootstrap. Ridge regression and lasso regression.
Introduction to classification problems. Bayes classifier and Naive Bayes. k-Nearest Neighbors (KNN), confusion matrix, and derived performance metrics.
Introduction to regression and classification decision trees. Tree pruning. Ensemble methods: bagging, random forests, and boosting.
Core Documentation
James, G., Witten, D., Hastie, T., Tibshirani, R. (2021). An Introduction to Statistical Learning: with Applications in R (2nd ed.). Springer.Type of evaluation
Oral examination. During the course, two midterm exams will be administered. teacher profile teaching materials
Review of multiple linear regression models. Basis functions for regression: polynomials, piecewise polynomials, and splines. Truncated basis functions and selection of the number of knots.
Indirect estimation of test error: Mallows' Cp, AIC, and BIC. Direct estimation of test error: validation set, LOOCV (Leave-One-Out Cross-Validation), K-fold cross-validation, and bootstrap. Ridge regression and lasso regression.
Introduction to classification problems. Bayes classifier and Naive Bayes. k-Nearest Neighbors (KNN), confusion matrix, and derived performance metrics.
Introduction to regression and classification decision trees. Tree pruning. Ensemble methods: bagging, random forests, and boosting.
Mutuazione: 21210514 Statistical learning in Economia e Gestione della Trasformazione Digitale LM-56 R FORTUNA FRANCESCA
Programme
Introduction to statistical learning. Training error and test error: implications of model flexibility in regression. Decomposition of test error. Bias–variance trade-off.Review of multiple linear regression models. Basis functions for regression: polynomials, piecewise polynomials, and splines. Truncated basis functions and selection of the number of knots.
Indirect estimation of test error: Mallows' Cp, AIC, and BIC. Direct estimation of test error: validation set, LOOCV (Leave-One-Out Cross-Validation), K-fold cross-validation, and bootstrap. Ridge regression and lasso regression.
Introduction to classification problems. Bayes classifier and Naive Bayes. k-Nearest Neighbors (KNN), confusion matrix, and derived performance metrics.
Introduction to regression and classification decision trees. Tree pruning. Ensemble methods: bagging, random forests, and boosting.
Core Documentation
James, G., Witten, D., Hastie, T., Tibshirani, R. (2021). An Introduction to Statistical Learning: with Applications in R (2nd ed.). Springer.Type of evaluation
Oral examination. During the course, two midterm exams will be administered.