Exploring the Generalizability of Spatio-Temporal Traffic Prediction: Meta-Modeling and an Analytic Framework

The Spatio-Temporal Traffic Prediction (STTP) problem is a classical problem with plenty of prior research efforts that benefit from traditional statistical learning and recent deep learning approaches. While STTP can refer to many real-world problems, most existing studies focus on quite specific applications, such as the prediction of taxi demand, ridesharing order, traffic speed, and so on. This hinders the STTP research as the approaches designed for different applications are hardly comparable, and thus how an application-driven approach can be generalized to other scenarios is unclear. To fill in this gap, this paper makes three efforts: (i) we propose an analytic framework, called STAnalytic, to qualitatively investigate STTP approaches regarding their design considerations on various spatial and temporal factors, aiming to make different application-driven approaches comparable; (ii) we design a spatio-temporal meta-model, called STMeta, which can flexibly integrate generalizable temporal and spatial knowledge identified by STAnalytic, (iii) we build an STTP benchmark platform including ten real-life datasets with five scenarios to quantitatively measure the generalizability of STTP approaches. In particular, we implement STMeta with different deep learning techniques, and STMeta demonstrates better generalizability than state-of-the-art approaches by achieving lower prediction error on average across all the datasets.

Automated summary

This paper proposes an analytic framework, STAnalytic, to explore the design considerations of spatio-temporal traffic prediction (STTP) approaches based on various spatial and temporal factors, aiming to achieve comparability among different application-driven approaches. Additionally, a flexible spatio-temporal meta-model, STMeta, is designed to integrate generalizable temporal and spatial knowledge identified by STAnalytic. Furthermore, a benchmark platform for STTP is built, consisting of ten real-life datasets and five scenarios, to quantitatively evaluate the generalizability of STTP approaches. Results show that STMeta using different deep learning techniques demonstrates superior generalizability with lower average prediction error across all datasets compared to existing methods.