Unlocking the Potential of Linear Networks for Irregular Multivariate Time Series Forecasting

2025-05-01

AI Summary

To address the challenges of modeling and forecasting irregular multivariate time series—characterized by non-uniform sampling intervals and high missing-data rates—this paper proposes Adaptive Linear Network (AiT). Methodologically, AiT introduces a temporal-adaptive linear layer that dynamically generates time-aware weights to explicitly model sampling irregularity; it further incorporates variable-specific semantic embeddings and a lightweight Transformer to jointly capture asynchronous cross-variable dependencies. Crucially, AiT achieves end-to-end accurate modeling of irregular multivariate time series for the first time while preserving the computational efficiency and interpretability of linear models. Extensive experiments on four benchmark datasets demonstrate that AiT achieves an average 11% improvement in forecasting accuracy and reduces inference latency by 52% compared to state-of-the-art methods, establishing new performance benchmarks for irregular time-series prediction.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Planning, Routing, and Scheduling: Planning with Language Models
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Time series forecasting holds significant importance across various industries, including finance, transportation, energy, healthcare, and climate. Despite the widespread use of linear networks due to their low computational cost and effectiveness in modeling temporal dependencies, most existing research has concentrated on regularly sampled and fully observed multivariate time series. However, in practice, we frequently encounter irregular multivariate time series characterized by variable sampling intervals and missing values. The inherent intra-series inconsistency and inter-series asynchrony in such data hinder effective modeling and forecasting with traditional linear networks relying on static weights. To tackle these challenges, this paper introduces a novel model named AiT. AiT utilizes an adaptive linear network capable of dynamically adjusting weights according to observation time points to address intra-series inconsistency, thereby enhancing the accuracy of temporal dependencies modeling. Furthermore, by incorporating the Transformer module on variable semantics embeddings, AiT efficiently captures variable correlations, avoiding the challenge of inter-series asynchrony. Comprehensive experiments across four benchmark datasets demonstrate the superiority of AiT, improving prediction accuracy by 11% and decreasing runtime by 52% compared to existing state-of-the-art methods.

Problem

• Addresses forecasting challenges in irregular multivariate time series
• Overcomes intra-series inconsistency with adaptive linear networks
• Resolves inter-series asynchrony via Transformer-based variable correlation

Innovation

• Adaptive linear network for dynamic weight adjustment
• Transformer module for variable correlation capture
• Improved accuracy and runtime efficiency