Machine Learning for Time Series
Preamble
Introduction
1. Defining Adequate Metrics for Structured Data
1.1 A Temporal Kernel for Time Series
1.2 Dynamic Time Warping
1.2.1 Constrained Dynamic Time Warping
1.2.2 DTW Alignment as an Adaptive Resampling Strategy
1.2.3 DTW with Global Invariances
1.3 Optimal Transport for Structured Data
2. Learning Sensible Representations for Time Series
2.1 Temporal Topic Models
2.2 Shapelet-based Representations and Convolutional Models
2.3 Early Classification of Time Series
Perspectives
Feedback welcome
Romain Tavenard's HDR thesis.
Powered by
Jupyter Book
.
.pdf