Title:         Machine Learning and Explainability
Instructor:     Damiano Verda, damiano.verda@rulex.ai
Credit points (CFU):    1
Lectures: 10 hours
Chapter Readings, Exercises, Literary Review and Final Project: 20 hours
Lectures Period: May 2025

Course Description and objectives

The goal of this course is to have a basic understanding of the main machine learning techniques, as well as to explore some of their applications. All the exercises will be shown (during the lectures) and carried out (for the final exam) by means of the self-coding Rulex Platform, so programming skills are not a requirement to attend the course (even if, of course, having this kind of skill and experience can be a plus).

The course aims to show that the scientific knowledge provided by these sectors is crucial to solve typical problems in finance, such as:

• knowing how to predict the trend (positive/negative) of prices in the energy trading market for athe considered geographical area (supervised learning, classification)
• knowing how to predict the expected price of a stock/asset (supervised learning, regression)
• knowing how to group stocks according to their market behavior (unsupervised learning, clustering)
  knowing how to extract associations between the price trends of different stocks/assets in the market (unsupervised learning, association rule mining) 
• assess and improve data quality
• evaluate and select the correct level of interpretability of the generated models, so that they can be understood and not acritically applied 

Prerequisites

A specific background is not strictly required for the reader, although basic notions of maths and statistics are recommended.

Course Materials

Lecture slides and exercises are provided to the students.

Assessment

Students will be assessed in the final exam, which will consist of completing a small machine learning flow by means of the Rulex Platform, with the help of the Rulex Academy e-learning website. The submission and the evaluation will also happen through the website.

Lecture Schedule 

Detailed program
Lecture I)  Supervised Machine Learning
Lecture: Wednesday 7th May 2025, 15-18 (3 hours).
 

Financial models discussed in the course are developed in Rulex Platform, as a result a brief tutorial of the software is provided to the students. Together with that, for the duration of the course, the students will have free access to the Rulex Academy e-learning portal. 
Key topics
•    Introduction to Machine Learning
•    Machine Learning Taxonomy
•    Regression
•    Feature Selection
•    Regression  On Time Series
•    Classification
•    Rule-Based Models
•    Rulex Example: Classification problem in energy trading

Study Materials

Rulex Academy: AI & Algorithms Learning Path (RC-AI)

More specifically, the following modules:
Increasing Customer Loyalty in Rulex Factory (Classification)
This course provides an example of how to use a classification task to understand and improve customer loyalty, and how we can use Rulex's proprietary classification algorithm (LLM) to solve the problem.

Predicting House Prices in Rulex Factory (Regression)
This course provides an example of how regression can be used to make quantitative predictions, including the theory behind these operations, and a hands-on real-life scenario to put the theory into practice.

References and further readings
Hastie, Trevor, Robert Tibshirani, and Jerome Friedman. "The elements of statistical learning: data mining, inference, and prediction." (2017).
Nasteski, Vladimir. "An overview of the supervised machine learning methods." Horizons. b 4.51-62 (2017): 56.
 

Lecture II) Unsupervised Machine Learning
Lecture: Wednesday 14th May 2025, 15-18 (3 hours).

Key topics
•    Introduction to Unsupervised Learning
•    Clustering
•    Tuning the number of clusters
•    Rule Clustering
•    Anomaly Detection
•    Frequent Itemsets and Association Rule Mining
•    Dimensionality Reduction
•    Rulex Example: clustering of stocks according to their historical behavior

Study Materials

Rulex Academy: AI & Algorithms Learning Path (RC-AI)

More specifically, the following modules: 
Targeting Marketing Campaigns (Clustering)
This course illustrates a real-life business example of how clustering can be used to extract value from data and create logical groupings, explains the theory behind these operations, and provides hands-on exercises to put the theory into practice.
Recommending Items in a Store with Rulex Factory (Association)
What are recommender systems? Why are they such a game changer for companies like Amazon and Netflix? This course will help you understand what lies beyond recommendations and how you can build your own.

References and further readings
Hastie, Trevor, Robert Tibshirani, and Jerome Friedman. "The elements of statistical learning: data mining, inference, and prediction." (2017).
Celebi, M. Emre, and Kemal Aydin, eds. Unsupervised learning algorithms. Vol. 9. Cham: Springer, 2016.

Lecture III) Advanced Concepts and Applications: Data Quality and Interpretability 
Lecture: Tuesday 20th May, 14:30-18:30 (4 hours).

Key topics
•    Definition and relevance of Data Quality
•    Data Quality issues: some examples
•    Use Case: checking and improving data quality on a sample dataset
•    Explainability definition and taxonomy
•    Pre-Modeling Explainability
•    In-Modeling Explainability
•    Post-Modeling Explainability
•    Use Case: extracting and visualizing an interpretable model on a sample dataset

Study Materials

Rulex Academy:  
Working with the Rule Engine Task - Rule Master

This course is about writing rules in the correct syntax, setting up an external rule configuration file, using the Rule Engine task to apply rules to data, and analyzing the most common mistakes users make in the process.
 

Privacy by Design with AI 
Are you planning on implementing new solutions based on artificial intelligence and you have doubts about privacy implications? This course will set you on the right path.

What makes a good use case for digital decision automation?
 

Digital decisions automation is revolutionizing industries and functions. Experts agree that those who succeed will have significant benefits and competitive advantages. However, many projects still fail. Why? This brief course will show you some tips on how to choose the best projects.

References and further readings
Fan, Wenfei, and Floris Geerts. Foundations of data quality management. Springer Nature, 2022.
Batini, Carlo, et al. "Methodologies for data quality assessment and improvement." ACM computing surveys (CSUR) 41.3 (2009): 1-52.
Longo, Luca, et al. "Explainable Artificial Intelligence (XAI) 2.0: A manifesto of open challenges and interdisciplinary research directions." Information Fusion 106 (2024): 102301.
Rashid, Muhammad, et al. "Can I trust my anomaly detection system? A case study based on explainable AI." World Conference on Explainable Artificial Intelligence. Cham: Springer Nature Switzerland, 2024.