Random Projection Layers for Multidimensional Time Sires Forecasting

arXiv:2402.10487v1 Announce Type: new
Abstract: All-Multi-Layer Perceptron (all-MLP) mixer models have been shown to be effective for time series forecasting problems. However, when such a model is applied to high-dimensional time series (e.g., the time series in a spatial-temporal dataset), its performance is likely to degrade due to overfitting issues. In this paper, we propose an all-MLP time series forecasting architecture, referred to as RPMixer. Our method leverages the ensemble-like behavior of deep neural networks, where each individual block within …

abstract arxiv cs.ai cs.lg dataset forecasting layer mlp multidimensional overfitting perceptron performance projection random series spatial temporal time series time series forecasting type