Predictive Models for Time Series Analysis
Components, transformations, and predictive models for time series, from DTW-based kNN to kernel- and dictionary-based approaches.
| Programme | Ph.D. in AI for Society |
|---|---|
| Degree | PhD |
| Institution | University of Pisa |
| Academic year | 2025/2026 |
| Period | Semester 2 (30 Apr – 19 May 2026) |
| Timetable | 30/04/2026 – 19/05/2026 |
| Hours | 24 |
| Credits | 3 CFU |
| Level | Introductory (no or few prerequisites) |
| Exam | Project |
| Location | MS Teams (online) |
| Lecturers | Francesco Spinnato, Riccardo Guidotti |
| Contact | francesco.spinnato@unipi.it, riccardo.guidotti@unipi.it |
| Official page | course catalogue |
| Teams channel | join the channel |
Lectures
The full content of the course is available on the Teams channel linked above, after approval.
Each lecture is 4 hours. The notebooks expect the datasets in a data/ folder next to them: download data.zip.
| Date | Lecture | Slides / Code |
|---|---|---|
| 30/04/2026 | Introduction & Preprocessing | course intro · slides · notebook · exercise |
| 05/05/2026 | Distances, Approximation & Global Features | slides · notebook · exercise |
| 07/05/2026 | Classification & Regression, Part 1 | slides · notebook · exercise |
| 12/05/2026 | Classification & Regression, Part 2 | slides · notebook · exercise |
| 14/05/2026 | Forecasting | slides · notebook · exercise |
| 19/05/2026 | In-class Project |