GinAR: An End-To-End Multivariate Time Series Forecasting Model Suitable for Variable Missing

TL;DR

GinAR addresses multivariate time series forecasting under variable missing by replacing all fully connected layers in SRU with Interpolation Attention (IA) and Adaptive Graph Convolution (AGCN), enabling end-to-end recovery of missing variables and correction of spatial-temporal dependencies. The model uses GinAR cells that perform missing-variable induction and data-driven graph learning, aggregating through multiple layers before an MLP decoder forecasts future values. Across five real-world datasets and missing rates up to 90%, GinAR outperforms 11 baselines, demonstrating strong robustness to missing data and avoiding the error accumulation common in two-stage imputations. This approach advances practical MTSF in incomplete-history settings by jointly recovering full variable sets and predicting future trajectories in a single, unified framework.

Abstract

Multivariate time series forecasting (MTSF) is crucial for decision-making to precisely forecast the future values/trends, based on the complex relationships identified from historical observations of multiple sequences. Recently, Spatial-Temporal Graph Neural Networks (STGNNs) have gradually become the theme of MTSF model as their powerful capability in mining spatial-temporal dependencies, but almost of them heavily rely on the assumption of historical data integrity. In reality, due to factors such as data collector failures and time-consuming repairment, it is extremely challenging to collect the whole historical observations without missing any variable. In this case, STGNNs can only utilize a subset of normal variables and easily suffer from the incorrect spatial-temporal dependency modeling issue, resulting in the degradation of their forecasting performance. To address the problem, in this paper, we propose a novel Graph Interpolation Attention Recursive Network (named GinAR) to precisely model the spatial-temporal dependencies over the limited collected data for forecasting. In GinAR, it consists of two key components, that is, interpolation attention and adaptive graph convolution to take place of the fully connected layer of simple recursive units, and thus are capable of recovering all missing variables and reconstructing the correct spatial-temporal dependencies for recursively modeling of multivariate time series data, respectively. Extensive experiments conducted on five real-world datasets demonstrate that GinAR outperforms 11 SOTA baselines, and even when 90% of variables are missing, it can still accurately predict the future values of all variables.

Figures (7)

• Figure 1: The principle and examples of multivariate time series forecasting with variable missing. V1 to V5 represent different variables. Compared to the other two tasks, our task can only use historical observations of certain variables to predict the future values of all variables. The forecasting performance of TGCN declines as the missing rate increases.
• Figure 2: (a) The overall framework of GinAR. The GinAR layer adopts the RNN-based sequence framework and encodes historical observations of MTS with variable missing. The MLP-based decoder is used to predict future values of all variables. (b) The specific structure of the interpolation attention. (c) The specific structure of the adaptive graph convolution.
• Figure 3: The schematic diagram of IA. Blue represents normal variables. White represents missing variables. Yellow represents the variables after induction.
• Figure 4: The results of the ablation experiment.
• Figure 5: Hyperparameter experiment results (PEMS04).