MH-GIN: Multi-scale Heterogeneous Graph-based Imputation Network for AIS Data (Extended Version)
Hengyu Liu, Tianyi Li, Yuqiang He, Kristian Torp, Yushuai Li, Christian S. Jensen
TL;DR
MH-GIN introduces a novel framework for imputing missing AIS data by explicitly modeling multi-scale dependencies among heterogeneous attributes. It combines type-specific encoders, a hierarchical temporal feature extractor, and a two-stage multi-scale graph propagation that aligns within scales and fuses cross-scale information across attributes. The approach achieves substantial imputation improvements over state-of-the-art methods (average gains around 57%) while maintaining efficiency, demonstrated on two real AIS datasets. This work enables more accurate maritime safety and monitoring applications by recovering high-fidelity spatio-temporal, cyclical, continuous, and discrete attributes across multiple update rates.
Abstract
Location-tracking data from the Automatic Identification System, much of which is publicly available, plays a key role in a range of maritime safety and monitoring applications. However, the data suffers from missing values that hamper downstream applications. Imputing the missing values is challenging because the values of different heterogeneous attributes are updated at diverse rates, resulting in the occurrence of multi-scale dependencies among attributes. Existing imputation methods that assume similar update rates across attributes are unable to capture and exploit such dependencies, limiting their imputation accuracy. We propose MH-GIN, a Multi-scale Heterogeneous Graph-based Imputation Network that aims improve imputation accuracy by capturing multi-scale dependencies. Specifically, MH-GIN first extracts multi-scale temporal features for each attribute while preserving their intrinsic heterogeneous characteristics. Then, it constructs a multi-scale heterogeneous graph to explicitly model dependencies between heterogeneous attributes to enable more accurate imputation of missing values through graph propagation. Experimental results on two real-world datasets find that MH-GIN is capable of an average 57% reduction in imputation errors compared to state-of-the-art methods, while maintaining computational efficiency. The source code and implementation details of MH-GIN are publicly available https://github.com/hyLiu1994/MH-GIN.