Generating Individual Trajectories Using GPT-2 Trained from Scratch on Encoded Spatiotemporal Data

TL;DR

This paper introduces a method to generate individual daily trajectories by encoding geographical, temporal, environmental, and personal data into tokens and training a GPT-2 model from scratch to produce realistic movement sequences.

Contribution

The study presents a novel approach of encoding spatiotemporal and personal data into tokens and training GPT-2 from scratch for trajectory generation, integrating environmental and individual factors.

Findings

Successfully generated realistic individual trajectories.

Incorporated environmental and personal factors into trajectory modeling.

Demonstrated the model's ability to reflect diverse movement patterns.

Abstract

Following Mizuno, Fujimoto, and Ishikawa's research (Front. Phys. 2022), we transpose geographical coordinates expressed in latitude and longitude into distinctive location tokens that embody positions across varied spatial scales. We encapsulate an individual daily trajectory as a sequence of tokens by adding unique time interval tokens to the location tokens. Using the architecture of an autoregressive language model, GPT-2, this sequence of tokens is trained from scratch, allowing us to construct a deep learning model that sequentially generates an individual daily trajectory. Environmental factors such as meteorological conditions and individual attributes such as gender and age are symbolized by unique special tokens, and by training these tokens and trajectories on the GPT-2 architecture, we can generate trajectories that are influenced by both environmental factors and individual attributes.