Laurent Oudre

Machine Learning for Time Series (Master MVA)

Teaching material and outline of the course Machine Learning for Time Series (Master MVA) during 2020-2021.

Course description

In many application contexts (health, economy, advertising...), the data collected takes the form of time series. The fundamental challenge then consists in choosing an adapted representation, allowing to take into account the temporal information as well as possible.

Machine Learning for time series gathers a large number of unsupervised or supervised tasks such as prediction, classification, completion/interpolation, query by content/indexation, clustering, segmentation/change-point detection or anomaly detection. But in reality, most of work for a data scientist dealing with temporal data consists in a series of hidden tasks such as:

• Understand the data: know where they come from, how they were acquired, what are their characteristics, interact with domain-experts and understand their problems
• Improve the data: find accurate representation spaces where the events of interest can be seen, consolidate the data (denoising, detrending, detection/removal of outliers)
• Model the data: physical/statistical or expert-based models, simple, adaptive and interpretable models
• Extract information from the data: find repetitive patterns, features of interest, change-points

This course aims to provide an overview of ML techniques to study time series, in different tasks such as pattern extraction and recognition, anomaly detection, prediction, interpolation etc. The course will mostly focus on these often poorly-documented hidden tasks and introduce several recent ML methods that will help the future data scientist to mine, but above all to understand time series. The course will be widely illustrated on real data and problems from current challenges and will emphasize aspects related to data understanding and interpretation. Note that in its current form, the course will only briefly discuss Deep Learning algorithms.

Presentation slides are available here

Outline and planning

Lectures and tutorial sessions will take place every Friday (afternoon or morning) at ENS Paris Saclay in Room 1Z33 (except for the oral presentations). For the tutorial sessions, students are asked to bring their personal laptops.

15/01/2021 14:00 → 17:00 Zoom	Introduction
	Lecture 1: Pattern Recognition and Detection Problem statement Comparing time series Euclidean distance Normalized Euclidean distance Dynamic Time Warping (DTW) Detecting patterns in time series Euclidean distance DTW Learning patterns from time series Distance-based pattern extraction Dictionary-based pattern extraction
22/01/2021 09:00 → 12:00 Zoom	Lecture 2: Feature Extraction and Selection Problem statement Feature extraction Spectral features Statistical features Deep Learning features Other features Feature selection Unsupervised setting Supervised setting
29/01/2021 14:00 → 17:00 Zoom	Tutorial 1 on Lectures 1 & 2	N/A
05/02/2021 14:00 → 17:00 Zoom	Lecture 3: Models and Representation Learning Problem statement Standard models Sinusoidal model Trend+Seasonality model AR models (and variants) Hidden Markov model Representation learning Standard representations Notion of sparsity Sparse coding Dictionary learning
12/02/2021 14:00 → 17:00 Zoom	Lecture 4: Denoising, Detrending, Interpolation and Outlier Removal Problem statement Denoising Filtering Sparse approximations Low-rank approximations Other techniques Detrending Least-Square regression Other techniques Interpolation of missing samples Polynomial interpolation Low-rank interpolation Model-based interpolation Outlier removal Isolated samples Contiguous samples
19/02/2021 14:00 → 17:00 Zoom	Tutorial 2 on Lectures 3 & 4	N/A
26/02/2021 14:00 → 17:00 Zoom	Lecture 5: Change-Point and Anomaly Detection Problem statement Change point detection Dealing with non-stationary time series Problem statement Cost functions Search methods Finding the number of change points Anomaly detection Outlier detection Statistical methods Model-based methods Distance-based methods
05/03/2021 14:00 → 17:00 Zoom	Lecture 6: Multivariate Time Series Problem statement Models for multivariate time series Vector autoregressive models Multivariate dictionary learning Graph signal processing Concepts and definitions Graph Fourier Transform (GFT) Bandlimitedness and smoothness Graph filtering Graph learning
12/03/2021 14:00 → 17:00 Zoom	Tutorial 3 on Lectures 5 & 6	N/A
26/03/2021 14:00 → 17:00 Room 1X10	Oral presentations

Validation

• Tutorials (25%): commented Jupyter notebook due at the end of each tutorial session
• Mini-project (75%):
• Report (25%): PDF file, 5 pages
• Source code (25%): commented Jupyter notebook
• Oral presentation (25%): 10 min presentation

Additional ressources

Useful references
List of possible topics/projects