Simple Feedfoward Neural Networks are Almost All You Need for Time Series Forecasting
This work provides a strong, simplified baseline for time series forecasting across domains like finance and healthcare, though it is incremental in challenging the necessity of complex models.
The paper tackles the problem of time series forecasting by demonstrating that simple feedforward neural networks (SFNNs) can match or exceed state-of-the-art models like Transformers and GNNs in performance, while being simpler, smaller, faster, and more robust, with analysis showing univariate SFNNs are often sufficient and multivariate ones remain competitive.
Time series data are everywhere -- from finance to healthcare -- and each domain brings its own unique complexities and structures. While advanced models like Transformers and graph neural networks (GNNs) have gained popularity in time series forecasting, largely due to their success in tasks like language modeling, their added complexity is not always necessary. In our work, we show that simple feedforward neural networks (SFNNs) can achieve performance on par with, or even exceeding, these state-of-the-art models, while being simpler, smaller, faster, and more robust. Our analysis indicates that, in many cases, univariate SFNNs are sufficient, implying that modeling interactions between multiple series may offer only marginal benefits. Even when inter-series relationships are strong, a basic multivariate SFNN still delivers competitive results. We also examine some key design choices and offer guidelines on making informed decisions. Additionally, we critique existing benchmarking practices and propose an improved evaluation protocol. Although SFNNs may not be optimal for every situation (hence the ``almost'' in our title) they serve as a strong baseline that future time series forecasting methods should always be compared against.