High dimensional regression for regenerative time-series: an application to road traffic modeling

TL;DR

This paper introduces a high-dimensional regression model for daily regenerating time-series data to predict urban road traffic, combining L1-penalization with risk bounds, and compares favorably to neural networks.

Contribution

It develops a novel high-dimensional vector autoregressive model with L1-penalization for daily regenerating time-series, providing theoretical risk bounds and practical advantages in traffic prediction.

Findings

Competitive prediction accuracy compared to neural networks

Identifies key road network sections affecting traffic

Establishes theoretical excess risk bounds in high-dimensional setting

Abstract

A statistical predictive model in which a high-dimensional time-series regenerates at the end of each day is used to model road traffic. Due to the regeneration, prediction is based on a daily modeling using a vector autoregressive model that combines linearly the past observations of the day. Due to the high-dimension, the learning algorithm follows from an L1-penalization of the regression coefficients. Excess risk bounds are established under the high-dimensional framework in which the number of road sections goes to infinity with the number of observed days. Considering floating car data observed in an urban area, the approach is compared to state-of-the-art methods including neural networks. In addition of being highly competitive in terms of prediction, it enables the identification of the most determinant sections of the road network.