Probabilistic Time Series Forecasting with Structured Shape and Temporal Diversity

TL;DR

This paper introduces STRIPE, a probabilistic forecasting model for non-stationary time series that captures structured shape and temporal diversity, improving prediction accuracy and diversity control.

Contribution

STRIPE is a novel, model-agnostic approach that uses differentiable DPP kernels and iterative sampling to explicitly control diversity in probabilistic time series forecasting.

Findings

STRIPE outperforms baseline methods in synthetic datasets.

STRIPE achieves superior results on real datasets compared to state-of-the-art approaches.

The iterative sampling scheme effectively disentangles shape and time representations.

Abstract

Probabilistic forecasting consists in predicting a distribution of possible future outcomes. In this paper, we address this problem for non-stationary time series, which is very challenging yet crucially important. We introduce the STRIPE model for representing structured diversity based on shape and time features, ensuring both probable predictions while being sharp and accurate. STRIPE is agnostic to the forecasting model, and we equip it with a diversification mechanism relying on determinantal point processes (DPP). We introduce two DPP kernels for modeling diverse trajectories in terms of shape and time, which are both differentiable and proved to be positive semi-definite. To have an explicit control on the diversity structure, we also design an iterative sampling mechanism to disentangle shape and time representations in the latent space. Experiments carried out on synthetic datasets show that STRIPE significantly outperforms baseline methods for representing diversity, while maintaining accuracy of the forecasting model. We also highlight the relevance of the iterative sampling scheme and the importance to use different criteria for measuring quality and diversity. Finally, experiments on real datasets illustrate that STRIPE is able to outperform state-of-the-art probabilistic forecasting approaches in the best sample prediction.