95866 - INTRODUCTION TO STATISTICAL LEARNING

Learning outcomes

The course aims at providing students with some of the main concepts and tools of Statistical Learning for economic and financial applications. These include tools for supervised learning, such as parametric and non-parametric regression and classification models, and resampling methods for model selection and assessment, such as cross-validation and bootstrap. The course emphasizes practical aspects by illustrating the methodology through empirical applications using the R software.

Course contents

The following topics will be covered:

1. Introduction and Overview of Statistical Learning
2. Linear Regression as a Prediction Tool
3. Moving Beyond linear regression: K-Nearest Neighbors.
4. Classification: K-Nearest Neighbors, Logistic Regression, misclassification rate, ROC curve, AUC.
5. Resampling Methods for Model Assessment and Selection: Bias-Variance trade-off; Training set, Validation set, Test set; Cross-Validation and the Bootstrap.

Readings/Bibliography

James, Witten, Hastie and Tibshirani, An Introduction to Statistical Learning, Springer 2021 (second edition).

Lecture notes will be made available at the beginning of the course.

Teaching methods

The emphasis of the course is on empirical applications. For this reason, for each topic, after introducing the relevant theory, illustrative real data examples will be given using the R language. Attending classes is important and highly recommended in order to better develop practical skills.

Assessment methods

The final examination covers the contents of both Module 1 and Module 2 according to the format detailed on the Module 2 website

Teaching tools

The statistical software R will be used to perform empirical studies. Additional resources include the software R Markdown for dynamic documentation and data repositories, such as Kaggle, UCI Machine Learning and Google Finance.

Office hours

See the website of Paola Bortot