Deep Time Series Forecasting with Shape and Temporal Criteria - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue IEEE Transactions on Pattern Analysis and Machine Intelligence Année : 2022

Deep Time Series Forecasting with Shape and Temporal Criteria

Résumé

This paper addresses the problem of multi-step time series forecasting for non-stationary signals that can present sudden changes. Current state-of-the-art deep learning forecasting methods, often trained with variants of the MSE, lack the ability to provide sharp predictions in deterministic and probabilistic contexts. To handle these challenges, we propose to incorporate shape and temporal criteria in the training objective of deep models. We define shape and temporal similarities and dissimilarities, based on a smooth relaxation of Dynamic Time Warping (DTW) and Temporal Distortion Index (TDI), that enable to build differentiable loss functions and positive semi-definite (PSD) kernels. With these tools, we introduce DILATE (DIstortion Loss including shApe and TimE), a new objective for deterministic forecasting, that explicitly incorporates two terms supporting precise shape and temporal change detection. For probabilistic forecasting, we introduce STRIPE++ (Shape and Time diverRsIty in Probabilistic forEcasting), a framework for providing a set of sharp and diverse forecasts, where the structured shape and time diversity is enforced with a determinantal point process (DPP) diversity loss. Extensive experiments and ablations studies on synthetic and real-world datasets confirm the benefits of leveraging shape and time features in time series forecasting.
Fichier principal
Vignette du fichier
full_paper.pdf (8.98 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03588390 , version 1 (24-02-2022)

Identifiants

Citer

Vincent Le Guen, Nicolas Thome. Deep Time Series Forecasting with Shape and Temporal Criteria. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 45 (1), pp.342-355. ⟨10.1109/TPAMI.2022.3152862⟩. ⟨hal-03588390⟩
94 Consultations
92 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More