Ing Ind - Inf (Mag.)(ord. 270) - MI (473) AUTOMATION AND CONTROL ENGINEERING - INGEGNERIA DELL'AUTOMAZIONE
*
A
ZZZZ
052351 - MODEL IDENTIFICATION AND DATA ANALYSIS
Ing Ind - Inf (Mag.)(ord. 270) - MI (476) ELECTRONICS ENGINEERING - INGEGNERIA ELETTRONICA
*
A
ZZZZ
052351 - MODEL IDENTIFICATION AND DATA ANALYSIS
Ing Ind - Inf (Mag.)(ord. 270) - MI (486) ENGINEERING PHYSICS - INGEGNERIA FISICA
*
A
ZZZZ
052351 - MODEL IDENTIFICATION AND DATA ANALYSIS
Obiettivi dell'insegnamento
The goal of the course is to enable students to extract (“learn”) from measured data useful information for automation, systems and control applications. The course covers the fundamentals of data-driven learning, as well as a series of methods for linear and nonlinear static systems modeling, clustering, time series characterization and virtual sensing in dynamical systems. Each technique is treated both theoretically and practically through computer exercises with real-world data.
Risultati di apprendimento attesi
Through theoretical lectures and computer sessions, the students are expected to:
understand the fundamental problems, shared by applied statistics, machine learning and system identification, that can be encountered while extracting information from a finite set of observations taken from an unknown system;
learn the main tools to avoid the most common problems in data-driven modeling, e.g., overfitting;
be able to formulate learning problems corresponding to different applications;
understand a range of statistical learning algorithms along with their strengths and weaknesses;
be able to apply learning algorithms to solve application problems;
be able to read current research papers and understand the issues raised by current research in the field.
Argomenti trattati
The program of the course is the following.
Introduction to statistical learning: main definitions and comparison with similar disciplines
The mathematical foundations of learning
Feasibility of statistical learning
Approximation VS generalization
Learning of static models
Linear regression
Logistic regression
Linear classification
Overfitting: regularization and validation
Neural networks
Data-preprocessing
Input pre-processing and data-cleaning
Dimensionality reduction and feature selection
Clustering
Towards learning of dynamical models
Stochastic processes
Data-preprocessing for time-series
Virtual sensing: the Kalman filter
Prerequisiti
Familiarity with basic concepts of computer science (algorithms and complexity) and dynamical systems theory.
Mathematical maturity in linear algebra and probability theory.
Modalità di valutazione
The exam will be a written test in one of the official available dates. The exam will consist of both theoretical questions and simple (i.e., no computer-aided) exercises. The final marks will be awarded based on correctness of the answers, appropriateness of technical terminology and clarity of exposition. Specifically, the students will be asked to:
show their understanding of the main problems (and the corresponding countermeasures) in the field of data-driven modeling and learning in general;
select and describe the most appropriate learning tools for a given application;
use analytical methods to solve learning problems of moderate complexity.
Bibliografia
Simone Formentin, Lecture notes and Python exercise notebooksYaser S. Abu-Mostafa, Malik Magdon-Ismail, Hsuan-Tien Lin, Learning from data, Editore: AML Book, Anno edizione: 2012
Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani., An Introduction to Statistical Learning: With Applications in R, Editore: Springer, Anno edizione: 2013 http://www-bcf.usc.edu/~gareth/ISL/ Note:
Inglese