Seasonal Encoder-Decoder Architecture for Forecasting

TL;DR

This paper introduces a novel RNN encoder-decoder architecture designed to effectively model seasonal patterns in time series data, enabling accurate multi-step forecasting even with limited data across multiple sequences.

Contribution

The paper presents a new RNN architecture that captures seasonal correlations and supports multi-step forecasting, including a recursive procedure for multi-sequence modeling with scarce data.

Findings

Effective in modeling seasonal correlations

Accurate multi-step forecasting demonstrated

Useful for single and multiple sequence data

Abstract

Deep learning (DL) in general and Recurrent neural networks (RNNs) in particular have seen high success levels in sequence based applications. This paper pertains to RNNs for time series modelling and forecasting. We propose a novel RNN architecture capturing (stochastic) seasonal correlations intelligently while capable of accurate multi-step forecasting. It is motivated from the well-known encoder-decoder (ED) architecture and multiplicative seasonal auto-regressive model. It incorporates multi-step (multi-target) learning even in the presence (or absence) of exogenous inputs. It can be employed on single or multiple sequence data. For the multiple sequence case, we also propose a novel greedy recursive procedure to build (one or more) predictive models across sequences when per-sequence data is less. We demonstrate via extensive experiments the utility of our proposed architecture both in single sequence and multiple sequence scenarios.