SSDNet: State Space Decomposition Neural Network for Time Series Forecasting

TL;DR

SSDNet is a novel deep learning model that combines Transformers with state space models to produce accurate, interpretable probabilistic forecasts for time series, capturing trend and seasonality without Kalman filters.

Contribution

This paper introduces SSDNet, integrating Transformer architecture with state space models for improved, interpretable time series forecasting without Kalman filters.

Findings

Outperforms existing deep learning and statistical methods in accuracy

Provides meaningful trend and seasonality components

Demonstrates efficiency and speed on multiple datasets

Abstract

In this paper, we present SSDNet, a novel deep learning approach for time series forecasting. SSDNet combines the Transformer architecture with state space models to provide probabilistic and interpretable forecasts, including trend and seasonality components and previous time steps important for the prediction. The Transformer architecture is used to learn the temporal patterns and estimate the parameters of the state space model directly and efficiently, without the need for Kalman filters. We comprehensively evaluate the performance of SSDNet on five data sets, showing that SSDNet is an effective method in terms of accuracy and speed, outperforming state-of-the-art deep learning and statistical methods, and able to provide meaningful trend and seasonality components.