96143 - METODI E MODELLI DI DATA ANALYTICS M

Course Unit Page

SDGs

This teaching activity contributes to the achievement of the Sustainable Development Goals of the UN 2030 Agenda.

Industry, innovation and infrastructure

Academic Year 2022/2023

Learning outcomes

Knowledge of the main concepts, techiques and tools for the design and implementation of "data analytics" and "data valorization" processes. All the phases of the data management and analysis process will be described, from initial data acquisition and pre-processing, to generation of new kowledge through statistic and Machine Learning techniques, up to visualization and performance evaluation. The main enterprise applications of data analytics and the main technological scenarios in this context will be presented.

Course contents

The course is composed of two modules:

  1. Data Mining (Fabio Grandi)
  2. Artificial Intelligence Methods  (Andrea Borghesi)

 

DATA MINING

Requirements/Prior Knowledge

A prior knowledge and understanding of database systems and relational model is required to attend with profit this course. These notions are normally achieved by giving an exam of Databases or Information Systems.

Fluent spoken and written Italian is a necessary pre-requisite: all lectures and tutorials, and all study material will be in Italian.

Course Contents

  • introduction to data mining
  • associative rules
  • clustering algorithms
  • typologies of data
  • decision trees
  • statistical methods
  • neural networks
  • evaluation of the results
  • analysis of time series
  • outlier detection

 

ARTIFICIAL INTELLIGENCE METHODS

Requirements/Prior Knowledge

A base knowledge of optimization and linear and integer programming is required to attend with profit this course.
These notions are normally achieved by giving an exam of Operations Research or Algorithms for Decision and Resource Management Support.

Fluent spoken and written Italian is a necessary pre-requisite: all lectures and tutorials, and all study material will be in Italian.

Course Contents

  • constraint programming
  • deep learning
  • automatic selection and configuration of algorithms
  • predictive maintenance
  • outlier detection
  • time series
  • prescriptive maintenance
  • use cases

Readings/Bibliography

  • Course slides.
  • I.H. Witten, E.Frank, M.A. Hall, Data Mining: Practical Machine Learning Tools and Techniques (third edition), Morgan Kaufmann, 2011.

Recommended readings:

  • M. Berry, G. Linoff. Data mining techniques for marketing, sales, and customer support. John Wiley & Sons, 1997.
  • R. Roiger, M. Geatz, Data Mining: A Tutorial-Based Primer, McGraw-Hill, 2003.

 

ARTIFICIAL INTELLIGENCE METHODS

  • Course slides.
  • I. Goodfellow, Y. Bengio, A. Courville, Deep Learning book, MIT Press, 2016. https://www.deeplearningbook.org/
  • T. Rossi, P. van Beek, T. Walsh, Handbook of Constraint Programming, Elsevier, 2006.

Teaching methods

  • Classroom lectures and exercises are given with the help of slides (through PC+projector presentations). Lectures could be possibly delivered in online or mixed modality using the Teams platform.
  • The program will be integrated by seminars from enterprise consultants.

Assessment methods

The final exam aims at verifying the knowledge acquired by the student, with regard to the specific contents of each teaching module, of the main available tools for the analysis of very large quantities of data to be used for efficiently supporting the decision processes. Exames will be conducted in the presence or online, according to the health situation and provisions of the University. In the online case, the Zoom and EOL tools will be employed.

For the two modules, final examinations are carried out separately. Each module examination is made of a 1-hour written test to be done without the aid of books or written notes. For the Data Mining module, the test consists in two (out of three) open-answer questions concerning the main course topics. For the Artificial Intelligence Methods module, the test consists in two (out of three) open-answer questions concerning the main course topics. Each module test is passed if it receives a 18/30 score on a total score of 32/30. Further details will be communicated during the lectures and in the "notes" appearing in the exam sessions published on AlmaEsami.

In order to take an exam, registration through AlmaEsami is required within the assigned deadlines. Students who won't be able to register within the deadline are bound to promptly (and anyway before the official closing of the registration lists) notify the issue to the didactic secretariat. It will be faculty of the teacher to admit them to the test. The tests of the two component modules of a course can be taken in the same or in different scheduled sessions, in any order. Once the outcome of a test has been published, each student is given a week to communicate via email to the teacher his/her intention to refuse the mark obtained. The final exam grade of a course is the average of the scores obtained for the two component modules.

To obtain a passing grade, students are required to at least demonstrate a knowledge of the key concepts of the subject, acquired autonomous design skills, and a comprehensible use of technical language. Higher grades will be awarded to students who demonstrate an organic understanding of the subject and a clear and concise presentation of the contents, a high ability for problem solving, and consistent design capabilities. A failing grade will be awarded if the student shows knowledge gaps in key-concepts of the subject, inappropriate use of language, logic failures in the analysis of the subject, inadequate operational and design skills.

 


Teaching tools

  • Materials on the course topics are available on the Virtual platform.
  • Teams platform for remote teaching.
  • Weka application software for exercising.

Office hours

See the website of Fabio Grandi

See the website of Andrea Borghesi