Tag: multivariate time series

A hybrid machine learning architecture optimized with quantum-inspired algorithms forecast wholesale rice prices across 23 Indian states with high short-term ...

A meta-transformer model combining multi-head attention with meta-learning improves vehicle lane-changing intention recognition accuracy by 8.7 percentage points using minimal ...

Researchers have developed WTCrossformer, an innovative hybrid deep learning architecture that integrates wavelet transform convolution and two-stage attention to achieve ...

A dynamic graph neural network learned spatial relationships directly from sensor streams to forecast municipal water quality hours in advance ...

Researchers developed IBformer, an architecture that introduces inductive bias and specialized enhancement modules to tackle spatial and temporal heterogeneity across ...

A new contrastive learning method called MsRCL uses multi-scale patch decomposition and channel-aware attention to improve anomaly detection in multivariate ...

Researchers at East China Jiaotong University have developed DE-SVDD, a dual-encoder framework that separates reconstruction learning from one-class boundary modeling ...

A new open-source package called Varmapack simulates VARMA time series exactly from the first observation, eliminating burn-in while running hundreds ...

Researchers have developed a hybrid deep learning architecture combining graph neural networks, Transformers, and variational autoencoders to detect and explain ...

Researchers at the University of Tabriz have developed GCAT, a graph-convolutional adversarial transformer that detects anomalies in multivariate sensor data ...