TTCD:Transformer Integrated Temporal Causal Discovery from Non-Stationary Time Series Data

Short summary

TTCD is a transformer-integrated framework that discovers contemporaneous and lagged causal relationships in non-stationary time series. It uses reconstruction-guided signal distillation to filter noise and spurious correlations while preserving meaningful dependencies. Experiments show the approach outperforms state-of-the-art methods on diverse datasets.

• •Novel transformer-based approach to temporal causal discovery
• •Handles non-stationarity, nonlinearity, and noise robustly
• •Outperforms existing methods on synthetic, benchmark, and real-world datasets

Generated with AI, which can make mistakes.