# Comparing linear structure-based and data-driven latent spatial representations for sequence prediction

## TL;DR

This paper compares linear structure-based and data-driven latent spatial representations to improve sequence prediction in graph-supported time series, addressing the joint modeling of spatial and temporal dependencies.

## Contribution

It provides an empirical comparison of different linear spatial representations for GTS prediction across multiple datasets.

## Key findings

- Data-driven representations outperform structure-based ones in prediction accuracy.
- Linear representations simplify modeling of complex spatial dependencies.
- Results vary depending on dataset characteristics.

## Abstract

Predicting the future of Graph-supported Time Series (GTS) is a key challenge in many domains, such as climate monitoring, finance or neuroimaging. Yet it is a highly difficult problem as it requires to account jointly for time and graph (spatial) dependencies. To simplify this process, it is common to use a two-step procedure in which spatial and time dependencies are dealt with separately. In this paper, we are interested in comparing various linear spatial representations, namely structure-based ones and data-driven ones, in terms of how they help predict the future of GTS. To that end, we perform experiments with various datasets including spontaneous brain activity and raw videos.