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

• DOI: 10.48550/arxiv.2405.07117
• Publication Date: 2024-05-11
• Authors: Abishek Sriramulu; Christoph Bergmeir; Slawek Smyl
• Keywords: FOS: Computer and information sciences; Computer Science - Machine Learning; Artificial Intelligence (cs.AI); Computer Science - Artificial Intelligence; Machine Learning (cs.LG)

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