TripCast: Pre-training of Masked 2D Transformers for Trip Time Series Forecasting

TL;DR

TripCast introduces a novel 2D transformer-based pre-training approach for trip time series forecasting, effectively capturing the unique leading time property in tourism data and outperforming existing methods.

Contribution

The paper proposes TripCast, a pre-trained 2D transformer model that leverages masking and reconstruction for improved trip time series forecasting.

Findings

Outperforms state-of-the-art baselines in in-domain scenarios

Demonstrates strong scalability and transferability in out-domain scenarios

Effectively captures the leading time property in tourism data

Abstract

Deep learning and pre-trained models have shown great success in time series forecasting. However, in the tourism industry, time series data often exhibit a leading time property, presenting a 2D structure. This introduces unique challenges for forecasting in this sector. In this study, we propose a novel modelling paradigm, TripCast, which treats trip time series as 2D data and learns representations through masking and reconstruction processes. Pre-trained on large-scale real-world data, TripCast notably outperforms other state-of-the-art baselines in in-domain forecasting scenarios and demonstrates strong scalability and transferability in out-domain forecasting scenarios.