<ul><li>HyperIMTS is a Hypergraph neural network designed for forecasting Irregular Multivariate Time Series (IMTS) that have irregular time intervals within variables and unaligned observations across variables.</li><li>Existing IMTS models face challenges such as the need for padded samples or representing original samples via bipartite graphs or sets to learn temporal and variable dependencies separately.</li><li>HyperIMTS overcomes these limitations by converting observed values into nodes in a hypergraph interconnected by temporal and variable hyperedges, allowing for message passing among all observations and capturing variable dependencies in a time-adaptive way.</li><li>Experiments have shown that HyperIMTS achieves competitive performance in forecasting IMTS with low computational cost compared to other state-of-the-art models.</li></ul>

HyperIMTS: Hypergraph Neural Network for Irregular Multivariate Time Series Forecasting

Discover more