Insegnamento a.a. 2025-2026

The course is organized along the following themes:

1. Introduction (including elements of optimization)

2. Linear models 

3. High-dimensional linear models

4. Classification

5. Probabillistic machine learning

6. Network data and algorithms 

• understand basic optimization
• understand basic machine learning pipeline
• understand basic predictive algorithms
• disinguish between prediction and causal inference
• identify network structures
• analyze network data

• build predictive algorithms
• evaluate predictive performance
• carry out network analytics

• Lectures
• Practical Exercises
• Collaborative Works / Assignments

• class lectures on the basis of slides and interaction with the class on fundamental concepts
• class hands-on sessions using jupyter notebooks that involve problem solving and programming

The course is evaluated in three different ways that involve both exams and continuous evaluation. 

One is an individual final exam. It involves 31 multiple choice questions, each carries 1 point. The exam is open book, you can use your laptop and any program that is locally installed (including LLMs) but cannot use the WiFi as it is an individual exam. Details about format, content and rules will be circulated through blackboard. The exam questions are organized in levels from 1 to 3, from easiest to hardest. There are about 15 level 1 questions, 11 level 2 questions and 5 level 3. Level 1 does not involve any coding, level 2 usually does, level 3 almost always does. The levels also reflect the time that it takes to answer the question. 

The other two involve continuous evaluation and a bonus-malus system.

Bonus: 
After each theme a set of multiple choice questions will be posted as a worksheet. It is the type of questions (occassionally, the exact same questions) that are included in the exam. After a week that they have been posted, during the following lectures (main lectures, LI sessions or TA sessions) the professor might ask the class about the previous questions,  what is the right answer and how the student arrived to the answer. People who volunteer and answer successfully they will be awarded with 0.25 points, for the easier questions, and 0.4 for the harder ones. Each student can accumulate a maximum of 2 bonus points like that, which will be added to their exam mark.

Malus:
At the beginning of the semester the students will be organized in groups of 3-5 people (no group will be accepted with less than 3 or more than 5 members). Many classes (both main lectures and LI sessions) are hands-on and involve coding. The jupyter-notebooks that develop the material to be covered in the class will be distributed to the class a week in advance. The students are expected to have gone through the notebooks before the class. Each notebook will contain a small number of easy exercises. The students,working as a group, should have prepared the answer to these questions before the class. During the class, using random numbers, the professor will ask a group to share their solution. If the group has not done anything, they get -0.25 points. If they have made some progress but not arrived to the solution, they get -0.1 points. If they have a reasonable answer, they get 0. The maximum of malus points received in this way is -2, which will be subtracted from the bonus points. Any student that has not joined a group will automatically get -2 malus points.

Example 1: a student with exam mark 28, that received 1.5 bonus points and their group has -0.5 malus points, will get 28+1.5-0.5 = 29

Example 2: a student with exam mark 16, that received 0 bonus points and did not join any group, will get 16
 

There will be slides provided for the methodological part of the course and they will form an important part of the reading material for understanding the main concepts. In terms of book references, one is

https://press.princeton.edu/books/hardcover/9780691222271/quantitative-social-science

which is used in Data Analytics but will also be relevant here.

A more related textbook, available online, that can be consulted and used together with the course materials is:

https://www.statlearning.com/

Additionally, this book:

https://www.deeplearningbook.org/

which is also freely available online and can be consulted.