Boundary-Aware Periodicity-based Sparsification Strategy for Ultra-Long Time Series Forecasting

Yiying Bao, Hao Zhou, Chao Peng, Chenyang Xu, Shuo Shi, Kecheng Cai

Abstract: In various domains such as transportation, resource management, and weather forecasting, there is an urgent need for methods that can provide predictions over a sufficiently long time horizon to encompass the period required for decision-making and implementation. Compared to traditional time series forecasting, ultra-long time series forecasting requires enhancing the model’s ability to infer long-term series, while maintaining inference costs within an acceptable range. To address this challenge, we propose the Boundary-Aware Periodicity-based sparsification strategy for Ultra-Long time series forecasting (BAP-UL). The periodicity-based sparsification strategy is a general lightweight data sparsification framework that captures periodic features in time series and reorganizes inputs and outputs into shorter sub-sequences for model prediction. The boundary-aware method, combined with the bounded nature of time series, improves the model’s capability to predict extreme peaks and irregular time series by adjusting the prediction results. We conducted extensive experiments on benchmark datasets, and the BAP-UL model achieved nearly 90% state-of-the-art results under various experimental conditions. Moreover, the data sparsification method based on the periodicity, proposed in this paper, exhibits broad applicability. It enhances the upper limit of sequence length for mainstream time series forecasting models and achieves the state-of-the-art prediction results.

Primary Subject Area: [Generation] Multimedia Foundation Models

Relevance To Conference: Time is the fundamental dimension used in various multimedia data prediction tasks. The core contribution of this study is a method that can discover the periodicity of sequences and incorporate periodic features for time series prediction. It is an effective approach for analyzing and mining various types of multimedia data with periodic characteristics. The framework proposed in this paper possesses excellent compatibility for processing large-scale variable data. It organizes images, text, and audio into sequential information on the same time dimension. The model can analyze the interrelationships between different modalities, thereby promoting the development of various multimedia data prediction tasks that encompass the time dimension.

Submission Number: 5703