93050 - Supervised Statistical Learning

Learning outcomes

By the end of the course the student knows the fundamentals of the most important multivariate techniques to build supervised statistical models for predicting or estimating an output based on one or more inputs. The student is able to represent and organize knowledge about large-scale data collections, and to turn data into actionable knowledge.

Course contents

Part 0: Introduction to Supervised Statistical Learning

Part 1: Resampling methods

• Cross-Validation

Part 2: Classification

• Naive Bayes
• k-Nearest Neighbours
• Logistic Regression
• Linear Discriminant Analysis

Part 3: Dimension Reduction and Regularisation

Part 4: Tree-based methods

• Regression and Classification trees
• Bagging; Random Forests; Boosting

Part 5: Visualizing classification results

Readings/Bibliography

The primary text for the course:

• James, G., Witten, D., Hastie, T., & Tibshirani, R. (2021). An introduction to Statistical Learning. Second Edition. New York: Springer. ISBN: 978-1-0716-1417-4. E-book ISBN 978-1-0716-1418-1
  
  The book is freely available here:
  https://www.statlearning.com/
  

 

In addition, we will use:

• T. Hastie, R. Tibshirani, and J. Friedman (2001) The Elements of Statistical Learning: data mining, inference and prediction. Springer Verlag.
  Freely available at:
  https://hastie.su.domains/ElemStatLearn/

Teaching methods

Lectures complemented with practical sessions.

As concerns the teaching methods of this course unit, all students must attend Module 1, 2 [http://www.unibo.it/en/services-and-opportunities/health-and-assistance/health-and-safety/online-course-on-health-and-safety-in-study-and-internship-areas ] on Health and Safety online.

Assessment methods

The learning assessment is composed by a written/practical test. The test is aimed at assessing the student's ability to use the learned definitions, concepts and properties and in solving exercises.

The exam consists of 5-10 questions, both multiple choice and open, some of which to be solved in R. The final grade is out of thirty. Students that, despite having passed the exam, do not feel represented by the obtained result can ask to have an additional (optional) oral exam (within at most 5-7 days) that can change the grade by +/-3 points. Please note: the difficulty of the oral questions will be calibrated based on the written exam grade. Questions will cover theoretical concepts and exercises from the entire syllabus.

During the written exam, students can only use the cheat sheet that is provided on virtuale.unibo.it, containing references to R packages and functions.

Students cannot make use of the textbook, personal notes, artificial intelligence tools nor mobile phones (smart watch or similar electronic data storage or communication devices are not allowed either and must be switched off before taking the exam).

Cheating or the use of unauthorized device is strictly prohibited. Any violation will result in the annulment of the exam and will be reported to the appropriate Division or Campus authorities in accordance with Article 48 of the University’s Code of Ethics.

To take the exam, students must register via the AlmaEsami platform. Students who do not register will not be allowed to take the exam.

Exams can only be taken during official sessions. No exceptions.

 

Students with learning disorders and\or temporary or permanent disabilities: please, contact the office responsible (https://site.unibo.it/studenti-con-disabilita-e-dsa/en/for-students) as soon as possible so that they can propose acceptable adjustments. The request for adaptation must be submitted in advance (15 days before the exam date) to the lecturer, who will assess the appropriateness of the adjustments, taking into account the teaching objectives.

Teaching tools

The following material will be provided: slides of the lectures, exercises with solutions, mock exam.

Office hours

See the website of Laura Anderlucci