Think Globally, Act Locally: A Deep Neural Network Approach to High-Dimensional Time Series Forecasting

TL;DR

DeepGLO is a novel deep neural network model designed for high-dimensional time series forecasting, combining global pattern learning with local calibration to improve accuracy on datasets with millions of correlated series.

Contribution

The paper introduces DeepGLO, a hybrid global-local deep learning model that effectively handles high-dimensional, diverse time series without normalization, outperforming existing methods.

Findings

DeepGLO achieves over 25% improvement in WAPE on a large public dataset.

The model effectively captures both global patterns and local properties of time series.

DeepGLO outperforms state-of-the-art approaches in high-dimensional forecasting.

Abstract

Forecasting high-dimensional time series plays a crucial role in many applications such as demand forecasting and financial predictions. Modern datasets can have millions of correlated time-series that evolve together, i.e they are extremely high dimensional (one dimension for each individual time-series). There is a need for exploiting global patterns and coupling them with local calibration for better prediction. However, most recent deep learning approaches in the literature are one-dimensional, i.e, even though they are trained on the whole dataset, during prediction, the future forecast for a single dimension mainly depends on past values from the same dimension. In this paper, we seek to correct this deficiency and propose DeepGLO, a deep forecasting model which thinks globally and acts locally. In particular, DeepGLO is a hybrid model that combines a global matrix factorization model regularized by a temporal convolution network, along with another temporal network that can capture local properties of each time-series and associated covariates. Our model can be trained effectively on high-dimensional but diverse time series, where different time series can have vastly different scales, without a priori normalization or rescaling. Empirical results demonstrate that DeepGLO can outperform state-of-the-art approaches; for example, we see more than 25% improvement in WAPE over other methods on a public dataset that contains more than 100K-dimensional time series.