Context Neural Networks: A Scalable Multivariate Model for Time Series Forecasting

TL;DR

This paper introduces the Context Neural Network, a scalable multivariate model that efficiently incorporates real-time contextual information from neighboring time series to improve forecasting accuracy without high computational costs.

Contribution

It presents a novel linear complexity approach for integrating inter-series context into time series forecasting models, overcoming scalability issues of existing multivariate methods.

Findings

Enables real-time context integration with linear complexity

Improves forecasting accuracy over traditional global models

Maintains computational efficiency for large datasets

Abstract

Real-world time series often exhibit complex interdependencies that cannot be captured in isolation. Global models that model past data from multiple related time series globally while producing series-specific forecasts locally are now common. However, their forecasts for each individual series remain isolated, failing to account for the current state of its neighbouring series. Multivariate models like multivariate attention and graph neural networks can explicitly incorporate inter-series information, thus addressing the shortcomings of global models. However, these techniques exhibit quadratic complexity per timestep, limiting scalability. This paper introduces the Context Neural Network, an efficient linear complexity approach for augmenting time series models with relevant contextual insights from neighbouring time series without significant computational overhead. The proposed method enriches predictive models by providing the target series with real-time information from its neighbours, addressing the limitations of global models, yet remaining computationally tractable for large datasets.